INTER- LABORATORY VARIABILITY OF HAEMOGLOBIN
MEASUREMENTS OBTAINED FROM SELECTED CLINICAL
LABORATORIES IN KENYA
MANDANIA ESTHER WANGUI
MASTER OF SCIENCE
(Medical Laboratory Sciences)
JOMO KENYATTA UNIVERSITY OF
AGRICULTURE AND TECHNOLOGY
2016
Inter- Laboratory Variability of Haemoglobin Measurements Obtained From
Selected Clinical Laboratories in Kenya
Mandania Esther Wangui
A Thesis Submitted in Partial Fulfilment for the Degree of Master of Medical
Laboratory Sciences in the Jomo Kenyatta University of Agriculture and
Technology
2016
i i
DECLARATION
This thesis is my original work and has not been presented for a degree in any other
university.
Signature……………………………….. Date……………………………………
Esther Wangui Mandania
This thesis has been submitted for examination with our approval as university
supervisors.
Signature……………………………….. Date……………………………………
Dr. Juliette R. Ongus, Ph.D.
JKUAT, Kenya
Signature……………………………….. Date……………………………………
Serah N. Kaggia, MMed (Human Pathology)
JKUAT, Kenya
Signature……………………………….. Date……………………………………
Fatmah Abdallah MMed (Human Pathology)
UoN, Kenya
i i i
DEDICATION
I dedicate this thesis to my family, my brothers and sisters for their moral support and
encouragement during the course of this study.
iv
ACKNOWLEDGEMENT
My greatest thanksgiving goes to the Almighty God for seeing me through my
Masters programme.
I wish to express my profound gratitude to my supervisors Dr. Juliette Ongus, Dr.
Serah N. Kaggia both of the Department of Medical Laboratory Sciences, COHES,
JKUAT and Dr. Fatmah Abdalla of Unit of Haematology and Blood Transfusion,
School of Medicine, UON for their close supervision and the inspiration that made
this thesis a reality.
I am indebted to Dr. Jane Carter, Head of Clinical Services, AMREF for allowing part
of this research to be conducted in Central laboratory, AMREF, Nairobi and Mr.
Stephene Munene for technical support.
My sincere gratitude also goes to all my lecturers at the Department of Medical
Laboratory Sciences, COHES, JKUAT for their guidance and contribution to this
thesis either directly or indirectly. I am also grateful to Mr. Patrick Chiyo for his
statiscal support.
Finally, my heartfelt gratitude goes to my family and all my friends who gave me
moral support in one way or another.
v
Table of Contents
DEDICATION ...................................................................................................................................................... III
ACKNOWLEDGEMENT .................................................................................................................................. IV
TABLE OF CONTENTS........................................................ ERROR! BOOKMARK NOT DEFINED.
LIST OF TABLES................................................................................................................................................ IX
LIST OF FIGURES................................................................................................................................................ X
OPERATIONAL DEFINATIONS .................................................................................................................. XI
LIST OF APPENDICES ............................................................................................................................... XVII
LIST OF ABBREVIATIONS AND ACRONYMS................................................................................XVIII
ABSTRACT .......................................................................................................................................................... XX
CHAPTER ONE............................................................................................................................................. 1
INTRODUCTION ......................................................................................................................................... 1
1.1. Background Information ................................................................................................................. 1
1.2. Statement of the Problem................................................................................................................ 3
1.3. Justification.......................................................................................................................................... 4
1.4. Research Questions ........................................................................................................................... 5
1.5. Hypothesis ............................................................................................................................................ 5
1.5.1 Null Hypothesis ............................................................................................................................. 5
1.5.2 Alternative Hypothesis ................................................................................................................ 6
1.6. Objectives............................................................................................................................................. 6
1.6.1. General Objectives .................................................................................................................. 6
1.6.2. Specific Objectives .................................................................................................................. 6
vi
CHAPTER TWO…………………………………………………………………...7
LITERATURE REVIEW........................................................................................................................... 7
2.2. Variability of Hb measurements................................................................................................... 7
2.1. Impact of Physiologic factors on haemoglobin measurements .......................................... 8
2.1.1. Type of Blood Sample................................................................................................................. 8
2.1.2. Sample Site ..................................................................................................................................... 8
2.1.3. Time ................................................................................................................................................... 9
2.1.4. Body Position ................................................................................................................................. 9
2.2. Laboratory devices and their impact on variability of Hb measurements ....................10
2.3. Accuracy and precision of Hb measurements........................................................................12
2.4. Quality control in haemoglobinometry ....................................................................................14
2.5. External Quality Assessment .......................................................................................................15
CHAPTER THREE ....................................................................................................................................17
MATERIALS AND METHODS ...........................................................................................................17
3.1. Study design ......................................................................................................................................17
3.2. Study site ............................................................................................................................................17
3.3. Study Population..............................................................................................................................17
3.4. Sample size ........................................................................................................................................18
3.4.1. District hospital laboratories and Sub- district hospital laboratories....................19
3.4.2. Private laboratories ................................................................................................................20
3.4.3. Health centre laboratories ...................................................................................................20
3.4.4. Accredited laboratories ........................................................................................................20
3.4.5. National referral laboratories .............................................................................................20
vii
3.5. Sampling Method ............................................................................................................................21
3.5.1. Inclusion Criteria....................................................................................................................21
3.5.2. Exclusion criteria ...................................................................................................................22
3.6. Laboratory procedures ...................................................................................................................22
3.6.1. Preparation of EQA samples ..............................................................................................22
3.6.2. Preparation of haemolysate ................................................................................................22
3.6.3. Assigning value of Hb concentration to the EQA samples (target values) ........23
3.6.4. Sample handling and storage .............................................................................................24
3.6.5. Sample packaging and transportation .............................................................................24
3.6.6. Sample processing .................................................................................................................25
3.7. Data Management and Analysis .................................................................................................26
3.8. Ethical Considerations ...................................................................................................................26
CHAPTER FOUR ............................................................................................................................................... 28
RESULTS .............................................................................................................................................................. 28
4.1. The performance of participating laboratories in differentiating, low, normal and
high Hb measurements .................................................................................................................................29
4.2. The inter-laboratory variability of the Hb measurements ..................................................30
4.3. Comparison of the variability of Hb measurements among the various Hb
measurement methods/equipment ............................................................................................................33
4.3.1. Variation of Hb measurements due to analysers .........................................................33
4.3.2. Variation of Hb measurements due to methods...........................................................39
CHAPTER FIVE ................................................................................................................................................. 42
DISCUSSION, CONCLUSIONS AND RECOMMENDATIONS.......................................................... 42
5.1. Discussion ................................................................................................................................................... 42
viii
5.2. Conclusions................................................................................................................................................. 49
5.3. Recommendations................................................................................................................................... 50
REFERENCES ..................................................................................................................................................... 51
APPENDICES...................................................................................................................................................... 63
ix
LIST OF TABLES
Table 3-1: Categories of Clinical laboratories in Kenya....................................................,....18
Table 3-2: Summary of the sampled laboratories for EQA.....................................................21
Table 4-3: The Different Analyzers and the Number of Laboratories Using Each Analyser
Type.........................................................................................................................................28
Table 4-4: The Number and Percentage of Laboratories that Performed Well (Passed) or
Failed according to CLIA’ 88 Test Performance Criteria........................................................30
Table 4-5: The Inter-Laboratory Variability of Hb Measurements for Each Type of Hb
Analyser...................................................................................................................................32
Table 4-6: The Accuracy of Analysers in estimating Hb values across all levels of
measurement in comparison to the expected or mean of all reference values (6.2, 13.6, and
18.1) combined (i.e. 12.633)....................................................................................................34
Table 4-7: The accuracy of the most commonly used analysers in Kenya in estimating low Hb
value (6.2 g/dl).........................................................................................................................36
Table 4-8: The accuracy of the most commonly used analysers in Kenya in estimating normal
Hb value (13.6 g/dl).................................................................................................................37
Table 4-9: The accuracy of the most commonly used analysers in Kenya in estimating high
Hb value (18.1 g/dl).................................................................................................................38
Table 4-10: The resuts for the questionnaire...........................................................................41
x
LIST OF FIGURES
Figure 4-1: Coefficient of variation as a measure of inter-laboratory variation for the three Hb
test concentrations...................................................................................................................31
Figure 4-2: Percentage number of laboratories as per each method of equipment /analyser
operation..................................................................................................................................39
Figure 4-3: Coefficient of variation around the standard reference as a measure of reliability
of the different methods of Hb analyser operation...................................................................40
xi
OPERATIONAL DEFINATIONS
Accuracy---------
Allowable Total
Error(TAE)-----------
Analytic Variability----
Bias( inaccuracy)------
Calibration-------------
Is the degree of closeness of measurements of a quantity to
that quantity’s actual (true) value.
A quality requirement that sets a limit for combined
imprecision (random error) and bias (inaccuracy, or
systematic error) that are tolerable in a single measurement
or single test result to insure clinical usefulness.
Random error in a laboratory result whose magnitude
depends on the methodology or analyser used.
Bias is the difference (generally unknown) between a
laboratory's average value (over time) for a test item and
the average that would be achieved by the reference
laboratory if it undertook the same measurements on the
same test item.
It’s a process of testing and adjusting an instrument or test
system, to establish a correlation between the measurement
response and the concentration of the substance that is
being measured by the test procedure.
xii
Coefficient of
variation---------------
Commutability-------
Deviation index---------
External Quality
Assurance (EQA)------
A measure of variability or diversity associated with
random error or imprecision. SD shows how much
variation or dispersion there is from the mean (average or
other Expected value) during repeated measures. A small
SD indicates that data points tend to be very close to the
mean, whereas a large SD indicates that the data points are
spread over a wide range of values. SD is the square root of
a dataset’s variance.
Degree to which a material yields the same numerical
relationships between results of measurements by a given
set of measurement procedures, purporting to measure the
same quantity, as those between the expectations of the
relationships for the same procedures applied to those
types of material for which the procedures are intended.
A measure of how far a result differs from the mean value
as multiples of the standard deviation.
A programme which determines total testing performance
by comparing a laboratory or clinic’s test result (including
interpretation of results) to a known standard or to an
appropriate peer group mean generated from an inter-
xii i
Imprecision(random
error) ------------
Inter-laboratory
Comparability-----------
Inter-Individual
Biologic Variation-----
Inter-laboratory
variability------------
laboratory comparison in which multiple laboratories
measure the same sample using the same test methods,
reagents and controls.
Lack of repeatability or reproducibility of the same result;
represented by the standard deviation (in units of the test)
or coefficient of variation, (in units of percent).
The organisation, performance and evaluation of
calibration/test results for the same or similar item by two
or more laboratories in accordance with predetermined
conditions
Variation in the mean values when a quantity is measured
in specimens from different individuals due to differences
in homeostatic set point of the individuals
It’s the variability of test results from different laboratories
using the same test method and analysing the same test
material. It includes both within-laboratory and between-
laboratory components of variance. Inter-laboratory
coefficients of variation (CVs) are CVs calculated based on
the total variance of results for a given method, endpoint,
xiv
Internal
Quality Control---------
Intra-Individual
Biologic Variation-----
Intra-laboratory
Variability-----------
and sample type.
It involves the in-house procedures for continuous
monitoring of operations and systematic day-to-day
checking of the produced data to decide whether these are
reliable enough to be released. The procedures primarily
monitor the bias of data with the help of control samples
and the precision by means of duplicate analyses of test
samples and/or of control samples.
Variation in results when a quantity is measured in
different specimens from the same individual obtained over
a time span due to the imprecision of the measurement
procedure (metrological variability) as well as to the
rhythmic and random fluctuations of the quantity value
around a virtual homeostatic set point.
Is the variability of test results from the same laboratory
using the same test method and analysing the same test
material. Within-laboratory coefficients of variation (CVs)
are CVs calculated based on solely the within-laboratory
component of variance.
xv
Intra-Subject
Variability ------------
Matrix-------------------
Matrix bias-------------
Precision-------------
Quality
Control (QC)----------
Includes differences in test results due to both subject
fluctuations and test performance fluctuations.
The matrix of the specimen is defined as the totality of
components of a material system except the analyte.
The component of the observed difference due to non-
commutability between a method/material combination.
Is the closeness of agreement between independent test
results obtained under stipulated conditions.
Procedures which monitor analytical performance of
instruments and detect analytical error. QC typically refers
to use of quality control materials and analysis of resulting
control data
Repeatability------------
Degree of consensus between successive measurements
which have been done on the same sample with very
similar conditions (same analyser, same user, same
laboratory, same methods, same lot of reagents) in a very
short time (e.g. same day). Often is symbolized as sr.
xvi
Reproducibility---------
Standard
Deviation (SD)----------
Trueness------------------
Is the degree of consensus between successive
measurements achieved on the same sample with different
conditions (e.g. different analyzer, different user, different
lot of reagents) in a long time. Can be either intra-
laboratory or inter-laboratory, and is symbolized as sR.
A measurement of imprecision (random error), biologic
variation, or other variability in a population;
mathematically, CV is standard deviation divided by the
mean and expressed as a percentage.
Closeness of agreement between the average value
obtained from a large set of test results and an accepted
reference value
xvii
LIST OF APPENDICES
APPENDIX I: Procedure for Haemolysate Preparation..........................................56
APPENDIX II: Standard Operating Procedure for Triple Packaging
System………………………………………………………………………….…..58
APPENDIX III: Instructions for the Participating Laboratories................................60
APPENDIX IV: Questionnaire...................................................................................62
APPENDIX V: Worksheet..........................................................................................65
APPENDIX VI: Ethical Clearance.............................................................................73
APPENDIX VII: Informed Consent Form for Laboratory Management...................68
xvii i
LIST OF ABBREVIATIONS AND ACRONYMS
AMREF African Medical Research Foundation
ANOVA Analysis Of Variance
BA Bland Altman
BGAs Blood gas analysers
CBC Complete blood count
CLIA Clinical Laboratory Improvement Amendments
CPD continuing professional development
CV Coefficient of Variation
DI Deviation Index
EA-REQAS East African Regional External Quality Assessment Scheme
EDTA Ethylene diamine tetra acetic acid
EQA External Quality Assurance
EQAS External Quality Assessment Schemes
ERC Ethical Review Committee
HA Haematology Analysers
Hb Haemoglobin
HICN Hemiglobincyanide
ICSH The International Committee for Standardisation in Haematology
IQA Internal Quality Assurance
IQR Inter-quartile Range
ISO International Organization for Standardization
KENAS Kenya National Assurance Schemes
xix
Km Kilometre
KMLTTB Kenya Medical Laboratory Technicians and Technologists Board
KNBTS Kenya National Blood Transfusion Service
KNH Kenyatta National Hospital
Lab Laboratory
mL Millilitre
NaCl Sodium Chloride
NIQR Normalized Inter-quartile Range
PBS Peripheral blood smear
PCV Packed Cell Volume
POC Point- of- care
PT Proficiency Testing
RBC Red blood cells
RCC Regional Co-ordinating Centre
SD Standard Deviation
SOPs Standard Operating Procedures
TLC Total leukocyte count
tHb Total Haemoglobin
UON University of Nairobi
WBC White blood cells
WHO World Health Organisation
xx
ABSTRACT
Accurate determination of haemoglobin (Hb) concentration is a common element in
assessing the extent of anaemia and an important variable in directing transfusion therapy in patients. Laboratory measurements should be reproducible and consistent
from day to day as well as between laboratories so that comparable results will be obtained when a given specimen is tested in different laboratories. Variability in Hb measurements is caused by many factors such as laboratory error, numerous
physiologic, temporal and methodologic factors. Therefore this study was aimed at determining the interlaboratory variability of Hb measurements obtained from
selected clinical laboratories across Kenya. A total of 188 public and 105 private laboratories selected from a total of 21 out of 47 Kenyan Counties received three specimens with low (A), normal (B) and high (C) Hb concentrations for analysis, after
which their results were compared and evaluated for accuracy. Laboratory performance was assessed using the Clinical Laboratory Improvement Amendments
of 1988 (CLIA’88) criteria. Coefficient of Variation (CV) was calculated as a measure of interlaboratory variability while the accuracy of the analysers was evaluated using one-way analysis of variance (ANOVA). A total of 67.98%, 64% and 50.6%
laboratories gave accurate results for samples A, B and C respectively. The results generated by the Celltac, Humalyzer Junior, Medonic, Mindray, Colourimeter,
Hemocontrol and Sysmex analysers were not significantly different (P>0.05) from the reference values. However, the Diaspect and Sahli analysers underestimated the Hb readings, while Hemocue, Urit and Mission overestimated the Hb readings (P<0.05).
Interlaboratory variation of 33.3%, 25.1% and 29.4% for samples A, B and C was recorded irrespective of the analyser a laboratory used. The CV for the automated,
semi-automated and manual methods was 7.08%, 7.04% and 34.26% respectively. The interlaboratory variation in Hb measurements resulted from variation in methodologies and types of analysers. Regular participation in External Quality
Assessment Schemes (EQAS) is essential in order to achieve interlaboratory comparability of Hb results. Laboratories should embrace automation and gradually
replace manual with automated methods, which are more accurate and reliable.
1
CHAPTER ONE
INTRODUCTION
1.1. Background Information
Accurate determination of Hb concentration is a common element in assessing the
extent of anaemia and an important variable in directing transfusion therapy in
patients. This decision should be made based on accurate, reliable and timely
laboratory tests. The detection of anaemia based on Hb concentration in venous or
capillary blood is used to estimate the prevalence of anaemia in populations, allocate
resources to programmes, and target intervention programmes to vulnerable groups;
it is also used to screen individuals for participation in programmes and to evaluate
response to interventions (Morris et al., 1999). The Hb test is used to indirectly asses
the oxygen carrying capacity of blood thus making it an important aid in
establishing the presence of anaemia and treating anaemia. The measurement of
blood Hb is one of the most common routine clinical laboratory tests. Hb
determination may be requested as an individual test or as part of a complete blood
count (CBC), which may be performed using either capillary or venous blood. The
Hb test is precise, simple to perform, easily standardised and may be performed
either manually or by using automated Hb analysers (Barbara, Anna & Norma,
2000).
The gold standard for assessing Hb concentration is the direct cyanmethaemoglobin
method (Burger and Pierre-Louis, 2003) and the International Committee for
Standardisation in Haematology (ICSH) recommends the Drabkin’s method as the
2
standard method for determining the Hb concentration of whole blood (WHO,
2008). Lately, new continuous, non-invasive methods of measuring Hb have been
introduced into the clinical environment whose accuracy may vary substantially
from invasive Hb measurements. This variability in measurements is caused by
many factors such as laboratory error, numerous physiologic, temporal and
methodologic factors (Lauren, 2013).
Despite the variability of Hb measurements, clinicians rely on accurate Hb values to
make decisions on transfusion and management of patients. Previous studies
conducted in other parts of the world have shown variation of Hb estimation
between laboratories (Blerk et al., 2007; Bilto, 1999; Renu et al., 2007).
The World Health Organisation (WHO) EQA for haematology assesses the
participating laboratories to correctly quantify the Hb level and number of white
blood cells and platelets (World Health Organisation, 2007). Patient Safety
Monitoring in International Laboratories (JHU-SMILE) remotely monitors
approximately 165 international laboratories and clinics in 22 countries, 84 of which
are in sub-saharan Africa. It is well known that laboratories in sub-saharan African
countries face many challenges to provision of quality results (WHO, 2008; Frean et
al., 2012). A 40 month study of the EQA performance of the Sub-Saharan African
JHU-SMILE laboratories reported that there was 0.6% failure rate for haemoglobin
in that cohort of laboratories (Amukele et al., 2012).
In East Africa, AMREF offers a number of programmes to address laboratory
quality, including continuing professional development, competency assessments,
blinded re-checking of test materials, regular structured support supervision and an
3
East African Regional External Quality Assessment Scheme (EA-REQAS). EA-
REQAS was established in 2003 when the four Ministries of Health of East Africa
(Kenya, Uganda, Mainland Tanzania, and Zanzibar) agreed to share laboratory
standards and materials; develop standard documents targeting both clinicians and
laboratory staff; establish national quality assurance bodies; form an EA-REQAS
Committee to oversee activities; and appoint AMREF as the Regional Co-ordinating
Centre. Seven surveys have been submitted since 2007, with nearly 400
participating facilities currently. Reference laboratories were identified to produce
quality PT (Proficiency Testing) materials. There has been progressive improvement
in performance by a few facilities participating in all surveys and national policy on
laboratory methodologies has been influenced (3rd Regional Technical Meeting EA-
REQAS, 2010)
1.2. Statement of the Problem
Although quality assurance programmes play an integral part in clinical laboratories
of developed countries, these programmes have not been accorded the same degree
of importance in the laboratories of developing countries. In Kenya in particular,
only few laboratories have participated in such schemes implying that the Hb
measurements of majority of clinical laboratories may not be accurate and
comparable. In addition, the variability of the Hb measurements between clinical
laboratories of Kenya has not been quantified. In order to address this problem, this
study sought to asses the extent of inter-laboratory variations in Hb estimation in
Kenyan clinical laboratories and to determine if the discrepancies are potentially
large enough to affect decisions consequently impacting on patient treatment and
4
management. Further, the study sought to evaluate the accuracy of the most
commonly used Hb analysers in Kenyan clinical laboratories.
1.3. Justification
Systematic error or bias may affect the measurement of a sample such that the result
obtained is not always a perfect value but may be some distance from the true value
(Kaplan, Pesce, & Kazmierczak, 2003). Use of different methodologies, reagents
and assay conditions (e.g. temperature, reagent concentrations, detection methods,
wash steps etc) are the reasons for the poor agreement between analysers. To
quantify these differences, methods need to be validated using reference
methodologies to ensure accuracy and comparability of results across laboratories
(Fierdoz, 2012).
Laboratories must continuously strive to maintain and improve the quality of
laboratory results, which is a never-ending process. Using statistical methods,
laboratories can continually improve their performance and the reliability of patient
results as well. Quality assurance in haematology is intended to ensure the reliability
of the laboratory results. A quality assurance programme has two main aspects,
namely, internal quality control (IQC) and external quality assessment (EQA) (Sah,
Raj & Prakash, 1999). IQC and EQA play very crucial roles in ensuring reliability
of analytical results (Westgard, 2008). EQAS of clinical laboratories ensures
between-laboratory comparability of results as well as detecting bias (systematic
errors) and overall review on the IQC programme (Rippey & Williamson, 1988;
Savage 1989).
Poor selection of techniques, lack of essential equipment, lack of quality control
materials and quality assurance systems, personnel issues and shortages of supplies
5
are some of the challenges to providing quality services faced by laboratories in
developing countries (Carter et al., 2002). Significant exchange of information in
patient care, teaching and research is becoming more and more difficult due to the
steadily increasing variation in methods and equipment (Munich, 1982). It is
therefore important that interlaboratory variability of Hb measurements and
reliability of equipment/analysers and methods used for Hb measurements in
Kenyan clinical laboratories be assessed to enhance quality. Despite the existence of
the EQA programmes in Kenya, however, laboratory participation is still very low
and optional. Therefore, it is essential that studies such as this present one are
frequently carried out to help in the improvement of interlaboratory comparison.
1.4. Research Questions
1. Can the laboratories participating in EQA of Hb measurements accurately
differentiate normal, low and high Hb specimen?
2. Is the inter-laboratory variability of the Hb measurements in an external quality
assessment quantifiable?
3. Can the variability of Hb measurements among the various Hb measurement
methods/equipment be compared?
1.5. Hypothesis
1.5.1 Null Hypothesis
There is no difference in Hb measurements obtained from selected clinical
laboratories in Kenya.
6
1.5.2 Alternative Hypothesis
There is difference in Hb measurements obtained from selected clinical laboratories
in Kenya.
1.6. Objectives
1.6.1. General Objectives
To determine the inter-laboratory variability of Hb measurements obtained from
selected clinical laboratories in Kenya.
1.6.2. Specific Objectives
1. To determine the ability of participating laboratories to accurately differentiate
normal, low and high Hb specimens.
2. To quantify the inter-laboratory variability of the Hb measurements.
3. To compare the variability of Hb measurements among the various Hb
measurement methods/equipment.
7
CHAPTER TWO
LITERATURE REVIEW
2.2. Variability of Hb measurements
Both accuracy (how close the measurement is to the actual value) and precision
(how repeatable the measurement is) of “standard” laboratory measurements are
subject to numerous methodologic factors that affect them (Lauren, 2013). Various
methods of Hb determination have been used all through ((Barbara, Anna & Norma,
2000). In the 1950s numerous different methods were in use leading to variability of
the results. This led to several attempts being made to standardise the measurement
of the Hb concentration in human blood. The Haemiglobincyanide method was well
accepted, soon came into general use, and performed satisfactorily for many years.
Later, with automation of laboratory methods in laboratory medicine, various
methods came up. This led to the manual Haemiglobincyanide method being phased
out as a routine method and became the reference method with which the current
methods should agree (Zigilstra, 1997). Measurement of total haemoglobin (tHb)
concentration by an automated laboratory analyser or point-of-care (POC) analyser,
using a venous or capillary blood sample is the conventional method. ICSH has
recommended the Drabkin’s method as the standard method for determining the Hb
concentration of whole blood (Chen et al., 1992).
The clinical measurement of tHb has inherent variability. Inter-device variability is
one cause of variability, for example, CO-Oximeter and POC devices that are
commonly used to measure Hb have been shown to vary up to ±1.2 and ±1.3 g/dL,
8
respectively. Additionally, there are a variety of physiologic and methodologic
factors that can significantly influence Hb levels in the body (Masimo, 2009).
2.1. Impact of Physiologic factors on haemoglobin measurements
A variety of factors influence Hb measurement and Hb levels in the body; these
include both the laboratory devices and physiologic factors. Sources of Hb variation
within the body include, the type of blood sample, sample site, time the sample is
taken, and body position.low Hb concentration readings: False high Hb
concentration readings:
2.1.1. Type of Blood Sample
Laboratory devices are designed to allow sampling of either venous or arterial
blood. Hb measurement may vary based on whether arterial or venous blood is used.
arterial Hb measurements can be expected to be, on average, 0.7 – 1.0 g/dL less than
the Hb measurements derived from venous blood (Mokken et al., 1996 & Yang ZW
et al., 2001) The percentage of plasma concentration can vary from the arterial to
venous blood based on a variety of physiologic factors, despite the total amount of
circulating red blood cells and Hb remaining relatively constant whether in arterial
or venous blood. The amount of plasma concentration can be higher in arterial
blood, potentially leading to lower concentration of Hb (Masimo, 2009)
2.1.2. Sample Site
The site on the body from where blood is drawn can also affect Hb measurements.
Large discrepancies were found between the values obtained from capillary blood
samples from the left and right hands of the same women, with intrasubject standard
deviation of 0.8 g/dL and correlation of 0.7 (Morris et al., 1999). The wide limits of
9
agreement indicate that two samples from different fingers of the same person could
have Hb concentrations that differ by up to 2.0g/dL. Another study shows wide
variation in the Hb concentration of capillary blood samples obtained from different
fingers on the same individual at the same time. Intrapatient variability ranged as
high as 7% (Bouton et al., 1994)
2.1.3. Time
Even in stable patients, Hb measurement can vary significantly over time. In a study
of venous blood samples drawn from the same individuals on two different
occasions, within person variances could vary as much as 2.6 g/dL in males and 2.3
g/dL in females (Looker AC, et al., 1990 & Burger et al., 2004). In another study,
when Hb measurements were taken from the same individual on four (4) different
days consecutively, intrasubject variability was 7.0% and the standard deviation was
0.8 g/dL ((Bouton et al., 1994)
2.1.4. Body Position
The position of the body before and during the blood draw also affects Hb
measurements due to the normal composition of blood, interstitial fluid shifts, and
elevations of protein and white blood cells. Body position has a significant effect on
venous Hb measurements due to decreases in plasma volume on assuming an
upright position. Heart rate and blood pressure are higher when standing vs. sitting,
which induces the movement of intravascular fluid such as plasma into interstitial
compartments. This causes Hct and Hb levels to rise (haemoconcentration) and
plasma volume to decrease (Martin et al., 1997). Gore and colleagues showed a 6%
10
reduction on plasma volume with standing, which changed Hb up to 2 g/dL (Gore
CJ, et al., 1992). Moving from seated to standing positions for 20 minutes may
result in a change in Hb concentration by >1.0 g/dL (Daniel-Johnson et al., 2007).
The converse is also true, indicating that if patients who are ambulatory change
body position prior to the blood draw, they may require a period of equilibration.
2.2. Laboratory devices and their impact on variability of Hb measurements
Historically venous, arterial, or capillary blood samples have been used for Hb
measurement; POC devices that use capillary blood samples. However, recently new
continuous, non-invasive methods of measuring Hb have been introduced into the
clinical environment whose accuracy can be variable compared to laboratory Hb
measures.
As clinicians interpret laboratory measurements, they expect that the values would
not change significantly if consecutive samples were measured repeatedly on the
same laboratory device or on different laboratory devices (Lauren, 2013). According
to the International Organization for Standardization (ISO), laboratory error is
“failure of a planned action to be completed as intended, or use of a wrong plan to
achieve an aim, occurring at any part of the laboratory cycle, from ordering,
examinations to reporting results and appropriately interpreting and reacting to
them” (Plebani & Lippi, 2010).
Care providers should take responsibility from the time of receipt of a request for a
pathology test/investigation, to the time outcomes are communicated to the requester
(The Royal College of Pathologists of Australasia, 2010). Therefore efforts to
11
reduce errors should consider the total testing procedure and quality improvement
should be measurable in terms of improvement to patient safety and clinical care
outcome (Plebani, 2010; Plebani, 2009). The chance of inappropriate care due to
laboratory errors ranges from 6.4 % – 12 %, although up to 30 % of errors also
resulted in patient discomfort, an escalation of costs and unwarranted additional
testing (Lippi & Guidi, 2007). Laboratory error, physiologic, temporal and
methodologic factors are the causes of the variability in reported Haemoglobin
values (Lauren, 2013). Multiple variables, which include device calibration, sample
handling, and other sources of variation specific to the technology, affect the
accuracy of each of the different methodologies used for total Hb measurement
(Lauren, Stephanie & Erin, 2011).
Both intra-device and inter-device variability affect total Hb measurements. Intra-
device comparison is Hb variability from the same blood sample on the same device
while inter-device comparison is Hb variability from the same sample on different
devices (Gehring et al., 2007). Previous studies have shown that there is significant
inter-device and intra-device variation in the Hb measurements; in 2007, five (5)
different models of CO-Oximeters were used to evaluate both inter-device and intra-
device variation in Hb measurements. When the same blood sample was analysed on
2 identical devices to test intra-device variability, the standard deviation between
measurements ranged from 0.2 to 1.2 g/dL (Gehring et al., 2007). Both reference
devices and test devices produce and/or contain inherent errors (Bland and Altman,
1986). The tHb measurements using a venous sample can vary as much as 0.9 g/dL
12
between different laboratory analysers across the normal tHb measurement range,
using a reference calibrator (RNA Medical, 2011).
Despite the use of POC haematology analysers becoming more frequent in the last
decade it has been shown that POC devices for Hb have reduced accuracy compared
to laboratory devices. Device methods, size of the blood sample, and strong
potential for confounding elements with capillary blood are factors that affect POC
device accuracy. At normal Hb levels of 13-15 g/ dL, the Clinical Laboratory
Improvement Amendments (CLIA) specification variance is approximately 1.0 g/dL
while in the anaemic range of 10 g/dL, the target variance is 0.7g/ Dl (Masimo,
2009). Previous studies reveal a significantly larger difference in haemoglobin
measurement between POC and laboratory devices; Hb measurement from capillary
blood in POC devices varies between 0.5 – 2.3 g/dL from reference standards
(Gomez-Simon et al., 2007; Patel et al., 2007; De Louw et al., 2007; Argawal &
Heinz, 2001).
2.3. Accuracy and precision of Hb measurements
Inter-laboratory imprecision as assessed by proficiency testing comprises
components of intra-laboratory imprecision and, more importantly, inter-laboratory
dissimilarities. Greater precision of techniques usually indicates a higher degree of
system stability among sites. Performance specifications for dispersion of results
and limits of acceptability are usually based on a percentage difference and /or
absolute difference from the target value (Rej & Jenny, 1992).
13
The Clinical Laboratory Improvement Act of 1988 (CLIA, 1988) define fixed limit
goals in absolute terms or multiples of standard deviations for a particular analyte.
According to CLIA 88 proficiency testing criteria, acceptable analytical
performance for haemoglobin is target ± 7% (CLIA, 1988).
The desirable specifications for within-subject biological variation, between-subject
biological variation, Allowable Total Error, Imprecision, and Bias (inaccuracy), for
Hb derived from Intra- and Inter-Individual Biologic Variation are 2.8%, 6.6%,
4.1%, 1.4% and 1.8% respectively (Ricos et al., 1999).
Many studies that have been conducted in many parts of the world have quantified
the imprecision, and accuracy (bias) of Hb measurements using the standard
deviation, coefficient of variation and the deviation index. The EQAS in
haematology at National Institute of Health, Islamabad, Pakistan (IEQAS)
conducted a survey to assess individual values against consensus value (mean ±SD)
and deviation index (DI) from the mean, whereas coefficients of variation (CV)
were calculated for years 1996 to 2006. The results were expressed as percentage of
accurate versus inaccurate results, DI and CV. The laboratory achieved 87.74% of
values within acceptable limits for Hb, 72.03% for white blood count, 69.49% for
platelet and 77.03% for reticulocyte estimation. These results were satisfactory,
having DI values less than 3 for all four parameters. Results were varied among
individual components of the survey but the overall DI values lay fairly well within
the acceptable limits with a majority of the results having a DI value less than 3.
Those results having DI values less than 3 were further classified into three
categories; excellent (DI > -2.0 to < 1.0), satisfactory (DI 1.0 to < 2.0) and out of
acceptable limits (DI < -2.0 or > 2.0). Overall, the laboratory achieved 64.22 % of
14
values with excellent results, 15.24% of results were within acceptable limits and
graded as satisfactory whereas 20.52% of values showed a high deviation from the
reference range. Mean DI values ± standard deviation and CV for Hb, were
calculated for each year. DI + standard deviation (SD) for Hb ranged between -
0.6±0.7 to 2.4±1.3.The study concluded that Participation in EQAS is extremely
beneficial for the improvement of laboratory performance and quality of care
(Birjees et al., 2009).
2.4. Quality control in haemoglobinometry
Regular calibration of spectrophotometers and Hb analysers used for Hb assays, as
specified by the manufacturer, must be done in order to achieve reliable Hb values.
Daily run of appropriate control solutions or when patient samples are run, recording
of control results and maintenance of records is a requirement. In addition each
instrument used must have its own particular checks that must be performed
(Barbara, Anna & Norma, 2000). Use of calibration specimens and high quality
control Hb solutions; checking and calibration of equipment are crucial for proper
determination of Hb in the blood. The results of Hb determination in control
solutions obtained in autoanalysers and manually, are comparable (Pupkova et al.,
2002).
Since 1996 the Hb standard (lot number BK-5182-2, later renamed lot 19-1-B518A)
has been in use till in 2008 when a new lot of the haemiglobincyanide (HiCN (Fe),
HiCN) standard was (#19-1-B806) was released by the ICSH in conjunction with
Eurotrol, B.V. The Haemoglobincyanide standard is used for the standardisation and
calibration of whole blood Hb measurements on most haemoglobinometers and
15
automated blood cell counters. This new lot was produced following the same
methodology previously specified by ICSH. Later both the WHO Expert Committee
on Biological Standardisation (International Reference Reagent 98/708) and the
European Community Bureau of Reference endorsed this lot as an international
reference material for Hb (Davis & Jungerius, 2010).
2.5. External Quality Assessment
Currently, EQAS exist in the field of laboratory medicine in many countries. Most
of these are intended to assist individual laboratories to continuously monitor their
performance and to compare it with that of other laboratories, whereas others may
be mainly intended for accreditation or licensing purposes. Additionally, EQAS may
monitor the quality of the commercial analytical systems, reagents and test kits, and
they help manufacturers to achieve a better harmonization of results from the
different analytical techniques (Ricos et al., 1996).
The two main aims of EQAS are to set both the target values and limits for
acceptance.
Reference methods are used to assign the target values, but since only a few
schemes follow these principles, target values are derived from the statics of each
survey (Ricos et al., 1996). Continual participation in EQA is an effective means for
identifying and mitigating variables that influence the reliability of analytical assays
for predictive markers, thereby assisting in technical validation and standardization.
The concept of EQA for the national health laboratories network is useful, as it
identifies problems in the comparability of laboratory results and initiates a process
16
towards solving these problems thus improving the quality of service at the level of
each individual laboratory and the network level (Olafsdottir et al., 1994).
The implementation of a quality assurance policy in a developing country requires a
commitment from the government, the professional societies and the laboratory
workers. It is important to recognize that a policy towards improving health care
should include an external surveillance system for health laboratories. An EQA
scheme will have the greatest impact when it is linked to a quality assurance
programme, which also includes internal quality control as an equally important
component (Deom et al., 1999).
17
CHAPTER THREE
MATERIALS AND METHODS
3.1. Study design
This was a cross-sectional study using inter-laboratory survey to assess the extent of
inter-laboratory variability of Hb assays in the clinical laboratories of Kenya. The
study lasted a period of 6 months (Between February and July 2016).
3.2. Study site
This study covered all the different laboratory categories which included national
reference laboratories, accredited laboratories, district & sub-district laboratories,
health centre laboratories and private hospital laboratories (level A, B, C, D and E
respectively). The study covered a total of 21 out of 47 counties in Kenya namely;
Bomet, Embu, Isiolo, Kajiado, Kericho, Kiambu, Kirinyaga, Kisii, Kitui, Laikipia,
Machakos, Makueni, Meru, Muranga, Nairobi, Nakuru, Narok, Nyandarua, Nyeri,
Tharaka Nithi and Uasin Gishu. The main study site was the Central Laboratory,
African Medical and Research Foundation, Kenya Located in Nairobi, from where
the samples were prepared and distributed to all the participating laboratories.
3.3. Study Population
The study population consisted of both Public and private clinical laboratories. All
laboratory categories, that is, national reference laboratories, accredited, district &
sub-district, health centres and private hospital laboratories were considered. A total
of 827 privately-owned laboratories, 2 national reference laboratories, 4 accredited
18
laboratories, 136 district hospitals, 130 sub-district and 1040 health centre
laboratories formed the study population from which the sample size was calculated.
3.4. Sample size
The sample size was calculated, as shown below, for the different levels of
laboratories using the population for each level. The number of laboratories
countrywide in each laboratory category represents the population from which a
representative sample was drawn. The following are the populations for the different
laboratory groups as indicated in the master facility list and as reported by the
KMLTTB (Kisabei, 2013, pers. comm).
Table 3-1: Categories of Clinical laboratories in Kenya
CATEGORY OF LABORATORY NUMBER
National reference laboratories 2
Accredited laboratories 4
District hospital laboratories 136
Sub- district hospital laboratories 130
Health centre laboratories 1,040
Private laboratories (Includes mission hospital laboratories) 827
Using Cochran’s sample size formula for continuous data (Cochran, 1977);
t)2 * (s)2
no= ----------------- (d)2
Where t = value for selected alpha level of 0.05(.025 in each tail) = 1.96
Where s = estimate of standard deviation in the population = 2.333
19
(Estimate of variance deviation for 14 point scale calculated by using 14 [inclusive
range of scale] divided by 6 [number of standard deviations that include almost all
(approximately 98%) of the possible values in the range]).
Where d = acceptable margin of error for mean being estimated = 0.42 (Number of
points on primary scale * acceptable margin of error; points on primary scale = 14;
acceptable margin of error = 0.03 [error researcher is willing to accept]). Therefore d
= 14 * 0.03 = 0.42.
3.4.1. District hospital laboratories and Sub- district hospital
laboratories
Population = 136+130 =266 laboratories
t)2 * (s)2
no= ----------------- = (1.96)2 (2.333)2 (d)2 14 *0.03)2
=118
However, since this sample size exceeds 5% of the population (266 *.05=13.3),
Cochran’s (1977) correction formula was used to calculate the final sample size.
These calculations are as follows:
no (118) 118 n i = ---------------------------- = -------------------- = 81 (1 + no / Population) (1 + 118/266)
Where n0 = required return sample size according to Cochran’s formula= 118.
Where n1 = required return sample size because sample > 5% of population.
Sample size = 81(40 district and 41 sub-district hospital laboratories)
20
3.4.2. Private laboratories
Population -------------------------------------827
t)2 * (s)2
no= ----------------- = (1.96)2 (2.333)2 (d)2 14 *0.03)2
=118
Correction formula:
no (118) 118 n= --------------------------------------------- = ----------------------------- = 103
(1 + no / Population) (1 + 118/827)
3.4.3. Health centre laboratories
Population----------------------------1,040
t)2 * (s)2 no= ----------------- = (1.96)2 (2.333)2
(d)2 14 *0.03)2
=118
Correction formula:
no (118) 118 n= --------------------------------------------- = ----------------------------- = 105 (1 + no / Population) (1 + 118/1040)
3.4.4. Accredited laboratories
Population------------------4
Sample size-----------------3
3.4.5. National referral laboratories
Population ---------------2
21
Sample size--------------1
The total sample size selected was … 81+103+105+3+1 =293 laboratories
Table 3-2: Summary of the sampled laboratories for EQA
CATEGORY OF LABORATORY NUMBER
National reference laboratories 1
Accredited laboratories 3
District hospital laboratories 40
Sub- district hospital laboratories 41
Health centre laboratories 105
Private laboratories 103
TOTAL 293
3.5. Sampling Method
Cluster random sampling of the laboratories from the 21 counties was used. To
arrive at the laboratory, a two-stage cluster sampling method was used. The counties
represented the clusters from which the laboratories were randomly selected. From
each county, a minimum of 7 public and 5 privately owned laboratories were
sampled. All laboratory categories were represented in all the counties except for
national reference and accredited laboratories which were sampled from the counties
they are located. Laboratories were selected from the rural, semi urban and urban
settings.
3.5.1. Inclusion Criteria
Both public and privately-owned laboratories that perform Hb measurements as one
of their routine laboratory tests and gave free informed consent for participation
were included in the study. All methods of Hb determination being used in the
laboratories were applicable.
22
3.5.2. Exclusion criteria
Private laboratories, which are not registered by KMLTTB (Kenya Medical
Laboratory Technicians and Technologists Board), and all laboratories that are
further than 315 kilometres (km) from the Central Laboratory, AMREF, Kenya were
excluded from the study because of logistical challenges.
3.6. Laboratory procedures
3.6.1. Preparation of EQA samples
EQA samples with low, normal and high Hb values (samples A, B and C)
respectively were prepared at the Central Laboratory AMREF, Nairobi (EAREQAS,
operating under African Medical and Research Foundation, Kenya).
3.6.2. Preparation of haemolysate
Three EQA samples with low, normal and high Hb values (samples A, B and C)
respectively were prepared as follows:-
One pint (450 ml) of blood in a blood bag which had tested negative for Human
immunodeficiency virus, hepatitis B virus surface antigen, syphilis and hepatitis C
virus antibodies was obtained from the Kenya National Blood Transfusion Service
(KNBTS). Using a centrifuge tube, 40 ml of the blood was centrifuged to separate
plasma and Buffy coat aseptically. To the red blood cells deposit, 2-3 fold volume of
physiological saline (9 g/L Sodium Chloride)(NaCl) was added, mixed well and
centrifuged at 2000g for 5 minutes. The supernatant and any remaining buffy coat
were discarded completely. This saline wash was repeated 2 times to ensure
complete removal of plasma, white cells and platelets. To the washed cells, half its
23
volume of carbon tetrachloride (neat), 99%, was added and the mixture shaken
vigorously in a mechanical shaker (Vibrofix VF1, Janke and Kunkel Model) at
2500g/minute for one hour. The mixture was then stored at 4oC overnight thus
forming a semi-solid interface of lipid/cell debris between carbon tetrachloride and
lysate. The mixture was centrifuged at 2500g for 20 minutes and the upper lysate
layer was carefully pipetted out into a clean Winchester bottle. Sterility and stability
of the haemolysate was maintained by the addition of preservatives and broad-
spectrum antibiotics as follows: To each 70 ml of lysate, 30 ml of glycerol was
added followed by addition of 25-50 mg of penicillin and 25-50 mg of gentamicin
per 500 mL of material. To make haemolysate with lower Hb concentration, an
appropriate volume of 30 % (v/v) glycerol in 9 g/L NaCl was added to the stock and
mixed well using a roller for one hour. While stirring continuously, 1 mL aliquots
were dispensed aseptically into 2 mL sterile vials, capped, sealed and labelled
appropriately. The samples were assigned unique codes A, B and C for sample with
low, normal and high Hb concentration respectively. The samples were preserved at
2-80 C in the refrigerator awaiting dispatch. An elaborate procedure of how the
samples were prepared is annexed (Appendix I).
3.6.3. Assigning value of Hb concentration to the EQA samples (target
values)
The Hb concentration for the three samples was assigned using the reference method
i.e. by use of a standard haematology analyser (Sysmex XS-800i, Sysmex
Corporation, Kobe, Japan) as the gold standard. These were used as the
target/reference values against which the results of the participating laboratories
24
were compared. The target/reference values for the three samples were 6.2 g/dl, 13.6
g/dl and 18.1 g/dl for sample A, B and C respectively.
3.6.4. Sample handling and storage
In order to minimise pre-analytical errors before and after sending materials to the
laboratories strict measures regarding sample aliquoting and sample handling were
taken. When preparing aliquots only one vial was handled at a time to avoid
exposing the samples to room temperature for more than 10 minutes. Samples were
checked for any leakage, spillage or contamination and for correct labelling with
well sticking labels. They were then stored according to temperature requirements
until testing can be performed, that is at 2-80C.
3.6.5. Sample packaging and transportation
The sample package for each participating laboratory contained three samples
labelled sample A, sample B and sample C. The samples were placed in leak proof 2
ml plastic vials which were properly labelled with the unique code numbers. The
quality of the samples was maintained during transport by use of icepacks so as to
maintain temperatures of 2-80C. Samples were also secured during transport so that
there is no leakage, spillage or contamination. During transport to the laboratories a
triple packaging system was used which consisted of a leak-proof primary
receptacle, a leak-proof secondary packaging with sufficient absorbent material and
an outer packaging of adequate strength (Appendix II). An efficient and reliable
means of transport was used.
25
In order to ensure that all the necessary requirements were followed by the
participating laboratories, the samples were accompanied by a complete set of
instructions with respect to storage, handling and deadline for analysis. The
instruction sheet contained information about the number of samples, type of
samples and details of how the samples are labelled, how to mix samples before
testing and a deadline for analysing the samples. Laboratories were informed to
perform checks for the samples before testing such as the storage conditions, code
numbers, breakage/ leakage, sign of contamination and the temperature condition at
which the samples were received (Appendix III).
3.6.6. Sample processing
The laboratories were instructed to process the samples within two days after
delivery using their current analytical procedures, to process the samples in the same
way as routine samples and record the results in the worksheet provided. Samples
were analysed in duplicate by performing two assays with a difference of not more
than six hours between the assays.
Alongside the study, each laboratory also received a questionnaire so as to collect
data on analytical aspects related to Hb determination. Data on analytic aspects of
Hb determination included, methods of analysis used, reference ranges used by each
laboratory, control materials and reagents used for analysis, use of Standard
Operating Procedures and previous participation in EQAS. Demographic data
included the code (assigned by Principal Investigator), location and mailing address
of the participating laboratories. The laboratories were also required to state clearly,
by filling the questionnaire, the date the samples were received, date of analysis and
26
the type of equipment/method used. The results were collected from the
participating laboratories within one week after dispatch of the samples.
3.7. Data Management and Analysis
Data was entered in Ms excel worksheet, coded and edited using consistency
checks, checks for duplicate entries and range checks. Data was analysed using
XLSTAT statistical software (XLSTAT Version 2013.3.03). For inter-laboratory
variability, coefficient of variation (CV) was used. The higher the CV the greater the
dispersion or spread in values of the variable, whereas when the CV is lower, the
residuals relative to the predicted value are smaller. The variation was also assessed
using analysis of variance (ANOVA). Significant differences (P<0.001) between
means were assessed by ANOVA. For evaluation of performance of participating
laboratories, acceptable performance criteria given by the CLIA’ 88 was used.
Accuracy of results was analysed by calculating the difference (expressed as the
bias) between the Hb concentration provided by the participating laboratories and
target values. We tested for the effect of the Hb concentration on the proportion of
the laboratories that performed well based on CLIA’88 criterion using Chi-Square
analysis. Data presentation was done using tables and graphs.
3.8. Ethical Considerations
Approval to carry out the study was sought from the Jomo Kenyatta University of
Agriculture and Technology as well as Kenyatta National Hospital/University of
Nairobi Ethics and Research Committee (Ref: KNH – ERC/A/1) (Appendix VIII).
Consent was obtained from the laboratory managers of the participating laboratories
27
prior to enrolment in the study. Confidentiality was maintained by coding the
participating laboratories rather than using their names. The questionnaires and
worksheets were kept under lock and key so that only the Principal Investigator
could access them.
28
CHAPTER FOUR
RESULTS
A total of two hundred and ninety two laboratories responded. Twenty seven
different analysers were used across all the laboratories (Table 4-3).
Table 4-3: The Different Analyzers and the Number of Laboratories Using
Each Analyser Type
Analyzer Method of
Operation
Number of Hb
Measurements
Number of
Laboratories
ABX Micros
Automated
6
1
ACT Diff Beckman Coulter
Automated 18 3
BTS 305 Manual 12 2
Celltac Automated 174 29
Cera Check Manual 18 3
Colourimeter Semi- Automated 78 13
Coulter Counter Automated 24 4
Diaspect Manual 306 51
Drew Automated 6 1
Easy Mate Manual 12 2
Sahli Manual 144 24
Hb Meter Manual 12 2
Hemocontrol Manual 264 44
Hemocue Manual 384 64
Hichroma Automated 6 1
Humalyzer Junior Semi- Automated 30 5
Hybrid Automated 12 2
Kyrot Automated 6 1
Medonic Automated 30 5
Mindray Automated 60 10
Mission Manual 36 6
Pentra ES 60 Automated 6 1
RMS Automated 6 1
Stat Manual 12 2
29
Sysmex Automated 48 8
Urit Manual 30 5
Erma Automated 12 2
4.1. The performance of participating laboratories in differentiating, low,
normal and high Hb measurements
Based on the CLIA’88 acceptable performance criteria for haemoglobin, which is,
target ± 7%, a total of 61% of laboratories had acceptable performance across all
measurements. The analyses shown in table 4-4 revealed that laboratory
performance using CLIA’88 criteria declined with increase in the concentration of
the target Hb value: 68%, 64% and 51% of laboratories passing for the sample with
low (6.2g/dl), normal (13.6g/dl) and high (18.1g/dl) Hb values respectively. These
differences were statistically significant (p <0.001) for the three Hb test
concentrations for both reading 1 and reading 2 (Table 4-4).
30
Table 4-4: The Number and Percentage of Laboratories that Performed Well
(Passed) or Failed according to CLIA’ 88 Test Performance Criteria
Measurement Reading 1 Reading 2 Average
Passed Failed Percent
success
Passed Failed Percent
success
Passed Failed Percent
success
Low (A) 198 94 67.81 199 93 68.15 397 187 67.98
Normal (B) 191 101 65.41 183 109 62.67 374 210 64.04
High (C) 148 144 50.68 148 144 50.68 296 288 50.68
Chi square 21.16 19.50 40.30
P value P<0.001 P<0.001 P<0.001
4.2. The inter-laboratory variability of the Hb measurements
The overall inter-laboratory CV was 33.3% for sample A, 25.1% for sample B and
29.4% for sample C irrespective of the analyser a laboratory used (fig 4-1). When
inter-laboratory CV was calculated across laboratories using the same analyser, the
CV reduced to 5.1 % (Hemocontrol) to 41 % (Urit) for sample A, 2.2% (Celltac) to
35% (Diaspect) for sample B and 3.4 % (Medonic) to 42.6% (Diaspect) for sample
C (Table 4-5).
31
Figure 4-1: Coefficient of variation as a measure of inter-laboratory variation
for the three Hb test concentrations
32
Table 4-5: The Inter-Laboratory Variability of Hb Measurements for Each
Type of Hb Analyser
Mean ± 1 Standard Deviation CV (%)
Hb
Analyzer
n A B C A B C
Celltac 29 6.43±1.09 13.83±0.30 18.7±0.97 16.9 2.2 5.2
Colorimeter 13 6.64±1.18 13.57±1.79 17.39±1.72 17.8 13.2 9.9
Diaspect 51 4.06±0.68 9.61±3.36 9.37±3.99 16.8 35.0 42.6
Hemocontrol 44 6.25±0.32 13.96±0.51 18.86±0.84 5.1 3.6 4.4
Hemocue 64 8.37±2.37 16.16±3.09 21.14±4.10 28.3 19.1 19.4
Humalyzer
Junior
5 6.10±0.43 13.00±1.07 17.64±0.84 7.1 8.3 4.8
Medonic 5 6.84±1.56 14.22±1.85 18.54±0.63 22.9 13.0 3.4
Mindray 10 5.99±0.6 13.33±0.48 18.35±0.78 10.1 3.6 4.3
Mission 6 11.25±2.62 16.52±1.49 19.67±1.14 23.3 9.0 5.8
Sahli 24 5.80±1.83 10.44±2.55 13.2±2.89 31.6 24.4 21.9
Sysmex 8 6.23±0.52 13.43±0.73 18.25±1.07 8.3 5.5 5.8
Urit 5 8.44±3.47 17.36±4.98 21.76±3.07 41.1 28.7 14.1
Average 269 6.465+2.12 13.39+4.95 17.08+3.26 32.85 24.38 28.98
33
4.3. Comparison of the variability of Hb measurements among the various
Hb measurement methods/equipment
4.3.1. Variation of Hb measurements due to analysers
Table 4-6 shows the accuracy of Hb analysers in estimating low, normal and high
Hb measurements. The values obtained by the different analysers and methods were
compared to the mean of the reference values. The mean Hb measurements from the
various analysers were significantly different from the mean of the reference values
(F (27) = 17.382, P<0.001,). In spite of this, seventy eight percent (n=21/27) of the Hb
analysers produced mean Hb values that were not significantly different from the
mean of the reference values and only 22% (n=6 of 27) had mean values deviating
significantly from that expected mean of the reference values. Most of the common
analysers produced estimates of Hb that were consistently different or similar to the
reference values for sample A, B and C (Tables 4-7, 4-8 & 4-9). The exception was
Sahli and Mission. Sahli gave Hb estimates that were similar to the reference value
for sample A but gave results that were significantly different than the reference
values for sample B and C. Mission gave estimates that were significantly different
from sample A and B reference values but not for sample C reference value.
34
Table 4-6: The Accuracy of Analysers in estimating Hb values across all levels
of measurement in comparison to the expected or mean of all reference values
(6.2, 13.6, and 18.1) combined (i.e. 12.633)
Analyzer N Mean
Hb
Value Standard.
Error
T. Statistic P value
Intercept 162 12.63 12.633 0.403 31.353 < 0.001
ABX Micros 6 11.63 1.000 2.132 -0.469 0.639
Beckman
coulterAct Diff
18 12.92 0.283 1.274 0.222 0.824
BTS 305 12 10.53 2.100 1.534 -1.369 0.171
Celltac 174 12.99 0.356 0.560 0.636 0.525
Cera Check 18 10.89 1.739 1.274 -1.365 0.173
Colourimeter 78 12.6 -0.035 0.707 -0.049 0.961
Coulter
Counter
24 12.52 -0.117 1.122 -0.104 0.917
Diaspect 306 7.68 -4.951 0.498 -9.935 < 0.001
Drew 6 13.45 0.817 2.132 0.383 0.702
Easy mate 12 15.57 2.933 1.534 1.912 0.056
Hb Meter 12 16.76 4.125 1.534 2.688 0.007
Hemocontrol 264 13.05 0.416 0.512 0.812 0.417
Hemocue 384 15.23 2.593 0.480 5.398 < 0.001
Hichroma 6 12.75 0.117 2.132 0.055 0.956
Humalyzer
Junior
30 12.09 -0.540 1.019 -0.530 0.596
Hybrid 12 11.84 -0.792 1.534 -0.516 0.606
Kyrot 6 13.5 0.867 2.132 0.406 0.684
Medonic 30 13.13 0.497 1.019 0.487 0.626
Mindray 60 12.54 -0.090 0.775 -0.116 0.908
Mission 36 15.73 3.092 0.945 3.272 0.001
Pentra Es 60 6 12.77 0.133 2.132 0.063 0.950
35
RMS 6 12.45 -0.183 2.132 -0.086 0.931
Sahli 144 9.81 -2.826 0.587 -4.812 < 0.001
Stat 12 14.96 2.325 1.534 1.515 0.130
Sysmex 48 12.64 0.010 0.843 0.012 0.990
Urit 30 15.85 3.217 1.019 3.156 0.002
Erma 12 12.51 -0.125 1.534 -0.081 0.935
36
Table 4-7: The accuracy of the most commonly used analysers in Kenya in
estimating low Hb value (6.2 g/dl)
Analyzer Method
of
operation
N Mean Value Standard.
error
T
statistic
P value
Intercept 6.200 0.284 21.823 < 0.001
Celltac A 29 6.43 0.234 0.395 0.594 0.553
Colourimeter SA 13 6.64 0.438 0.498 0.880 0.380
Diaspect M 51 4.06 -2.137 0.351 -6.083 < 0.001
Hemocontrol M 44 6.25 0.048 0.361 0.132 0.895
Hemocue M 64 8.37 2.167 0.339 6.397 < 0.001
Humalyzer
Junior
SA 5 6.1 -0.100 0.719 -0.139 0.889
Medonic A 5 6.84 0.640 0.719 0.890 0.374
Mindray A 10 5.99 -0.210 0.546 -0.384 0.701
Mission M 6 11.25 5.050 0.666 7.579 < 0.001
Sahli M 24 5.8 -0.396 0.414 -0.956 0.340
Sysmex A 8 6.23 0.025 0.594 0.042 0.966
Urit M 5 8.44 2.240 0.719 3.117 0.002
KEY:
A – Automated
SA – Semi Automated
M - Manual
37
Table 4-8: The accuracy of the most commonly used analysers in Kenya in
estimating normal Hb value (13.6 g/dl)
Analyzer Method
of
operation
N Mean Value Standard
error
T
statistic
P value
Intercept 13.6 0.447 30.418 < 0.0001
Celltac A 29 13.83 0.231 0.621 0.372 0.71
Colourimeter SA 13 13.57 -0.031 0.784 -0.039 0.969
Diaspect M 51 9.61 -3.988 0.553 -7.213 < 0.0001
Hemocontrol M 44 13.96 0.357 0.568 0.628 0.53
Hemocue M 64 16.16 2.556 0.533 4.795 < 0.0001
Humalyzer SA 5 13 -0.6 1.131 -0.53 0.596
Medonic A 5 14.22 0.62 1.131 0.548 0.584
Mindray A 10 13.33 -0.27 0.86 -0.314 0.754
Mission M 6 16.52 2.917 1.049 2.782 0.006
Sahli M 24 10.44 -3.158 0.652 -4.846 < 0.0001
Sysmex A 8 13.43 -0.175 0.935 -0.187 0.852
Urit M 4 17.36 3.76 1.131 3.324 0.001
KEY:
A – Automated
SA – Semi Automated
M - Manual
38
Table 4-9: The accuracy of the most commonly used analysers in Kenya in
estimating high Hb value (18.1 g/dl)
Analyser Method
of
operation
N Mean Value Standard
error
T
statistic
P value
Intercept 18.100 0.540 33.498 < 0.001
Celltac A 29 18.7 0.600 0.751 0.799 0.425
Colourimeter SA 13 17.39 -0.708 0.948 -0.747 0.456
Diaspect M 51 9.37 -8.733 0.668 -13.069 < 0.001
Hemocontrol M 44 18.86 0.759 0.686 1.106 0.270
Hemocue M 64 21.14 3.038 0.644 4.714 < 0.001
Humalyzer Junior
SA 5 17.64 -0.460 1.367 -0.337 0.737
Medonic A 5 18.54 0.440 1.367 0.322 0.748
Mindray A 10
18.35 0.250 1.039 0.241 0.810
Mission M 6 19.67 1.567 1.267 1.236 0.217
Sahli M 24 13.2 -4.900 0.788 -6.221 < 0.001
Sysmex A 8 18.25 0.150 1.130 0.133 0.895
Urit M 5 22.63 4.525 1.504 3.008 0.003
KEY:
A – Automated
SA – Semi Automated
M - Manual
39
4.3.2. Variation of Hb measurements due to methods
A total of 74.32%, (n=217) of the laboratories used analysers/machines that
employed a manual method, 24% used an automated methods (n=70) and 1.7%,
(n=5) used semi-automated methods (fig 4-2). ANOVA comparisons revealed that
the mean Hb across the three methods of analyser/machine operation were not
significantly different from the mean of the reference values (F(3) =1.333, P=0.262,
mean of reference value =12.633). When we calculated the CV using the mean of
reference values, the CV for the automated and semi automated methods were
similar but was large for analysers using the manual method (CV for automated
analyzers =7.08%, CV for semi automated analyzers = 7.04% and CV for manual
analyzers =34.26% (fig 4-3).
Figure 4-2: Percentage number of laboratories as per each method of
equipment /analyser operation
40
Figure 4-3: Coefficient of variation around the standard reference as a measure
of reliability of the different methods of Hb analyser operation
4.4. Resuts for the questionnaire
Important analytical data regarding the practice of internal quality control in the
laboratories was as shown in table 4-10.
41
Table 4-10: The resuts for the questionnaire
S/No Question Number Of
Laboratories
Percentage
1. Used SOPs during testing 200 68.5 %
2. Previous participation in EQA 44 15%
3. Used quality control materials 131 45%
4. Reported the correct reference values for Hb 254 87%
5. Used manual methods 217 74.3%
6. Used automated methods 70 24%
7. Used semi-automated methods 5 1.7%
42
CHAPTER FIVE
DISCUSSION, CONCLUSIONS AND RECOMMENDATIONS
5.1. Discussion
As demonstrated in this study, 61% of laboratories met the CLIA’88 criterion for
samples A, B and C (Table 4-4). The ability of a laboratory to accurately determine
the Hb values of a sample depended on the type of analyser/method used and the
application and adherence to quality control practices by the laboratory. The results
in this study have demonstrated variation of Hb results when laboratories analyse
the same sample. Such results will have potential implications in the classification of
anaemic patients. For instance, assuming sample A, B and C was representative of
an anaemic subject, a healthy normal subject and a subject with higher Hb
concentration respectively. A total of 1.7 % (5 out of 292), 19.8 % (58 out of 292)
and 37.3 % (109 out of 292) laboratories respectively, would have misclassified
these cases.
The present study demonstrates inter-laboratory variation in Hb measurements for
the three Hb samples, which was slightly higher at low Hb concentration. Patients
usually obtain health care services in different clinical settings and therefore may
have their Hb levels measured in different laboratories using different methods that
may not be comparable. In this study the range of Hb results obtained for each
analyser/method varied widely. Confusion may arise when laboratories give
different results for the same sample. For instance, in the present study there was a
considerable high overall interlaboratory variation. However, when the analysers
43
were grouped according to the type of analyser, the variation of Hb results reduced
to a great extent. This suggests that both method/analyser bias and laboratory-
specific bias are the cause of the overall variability of interlaboratory Hb results.
Unfortunately, the interlaboratory variation may have consequences for clinical
practice depending on the case at hand. The results obtained in the present study are
similar to those found in a survey of Belgian hospitals to assess the reliability of Hb
measurements which reported interlaboratory variation ranged from 0.6 % (Roche)
to 6.7 % (IL) for sample 1 and from 2.0 % (Radiometer) to 4.5 % (IL) for sample 2
( Blerk et al., 2007). Also, a study carried out to assess the analytical quality of
Jordanian haematology laboratories in the routine haematological parameters
reported that there was considerable inter-laboratory variation in laboratories using
manual methods (Bilto, 1999).
In this study, some analysers gave results that are comparable to the reference values
while as others overestimated and others underestimated the values. An overall
negative bias was observed for Hb results obtained using Diaspect and Sahli.
Analytic bias can directly affect patient classification and clinical decision making
(Klee, 1995). In addition, underestimation and overestimation of Hb concentrations
could lead to unnecessary clinical interventions in patients. Some analysers gave
results that are comparable to the reference values while as others overestimated and
others underestimated the values.
The underestimation of Hb values by both Diaspect and Sahli appears to increase
with an increase in Hb concentration. These results supports findings from other
44
studies, which reported that Sahli method underestimated Hb values. According to
Patil, Thakare and Patil, (2013) Sahli underestimates the Hb in capillary blood by
0.62gm/dl and 1.12g/dl in venous blood. Kapil et al., (2002) compared HemoCue®
analyser with Sahli’s method and they found that Sahli’s method underestimates Hb
by 1.06 g/dl compared to HemoCue® analyser. Barduagni et al., (2003) compared
the performance of the Sahli and the colour scale methods in diagnosing anaemia in
school children and found that Sahli method gave low mean Hb level (11.4g/dl) and
had low specificity (39.0%), thus giving a high rate of false positive results. The
considerable variability in the Hb values obtained with the Sahli could be due to its
inbuilt errors, subjective visual colour comparison, inaccuracy in pipetting of blood,
fading of comparator after prolonged use and poor sensitivity and reliability.
Contrary to the results presented in this study, a study conducted to evaluate the
utility of the Diaspect as a point of care analyser reported that the Hb values
obtained with Diaspect system compared well with the reference analyser
(Robertson, Lewis, & Osei-Bimpong 2011). It is evident from the findings of this
study that both Diaspect and Sahli methods may classify healthy individuals as
anaemic.
On the other hand, Hemocue, Urit and Mission were consistently overestimating the
Hb values when compared to the reference. Overestimation of Hb values by
Hemocue and Urit appears to increase with an increase in Hb concentration with an
exception of Mission where the overestimation decreases with an increase in Hb
concentration. Hemocue, Urit, Mission showed a positive bias in that order of
increasing bias.
45
The reliability, precision and accuracy of hemocue is controversial. The results of
this study showed that hemocue overestimated Hb values by more than 2g/dl from
the reference and this overestimate appears to increase with increase in the actual Hb
value. The present study findings agree with Mohanram, Ramana-Rao & Sastry,
(2002) who observed that hemocue method overestimated Hb by more than 2g/dl.
Prakash et al., (1999), in their study on utility of hemocue in estimation of Hb
against standard blood cell counter reported that hemocue produced consistently
higher values of Hb than the blood cell counter and suggested that a constant
correction factor of 0.5g/dl should be subtracted from Hb estimates of hemocue.
Bhaskaram et al., (2003) also found that hemocue overestimated Hb values by 10%
- 15% when they compared it to cyanmethaemoglobin method in their study on
validation of Hb estimation using hemocue and this overestimate appeared to
increase with increase in the actual Hb value. However other studies such as those
conducted by Sari et al., (2001) to estimate the prevalence of anaemia among
Indonesian mothers, Von Schenck, Falkensson and Lundergerg, (1986) in evaluation
of hemocue and Rechner et al., (2002) on evaluation of hemocue compared with the
coulter STKS have found hemocue to be reliable and accurate and have supported
its use. Yet, a study conducted in Mexico reported that hemocue underestimated Hb
values (Neufeld et al., 2002). However, due to these contradicting findings on the
accuracy and reliability of hemocue, it is suggested that further studies need to be
done to validate the hemocue haemoglobinometer.
The findings in this study about Urit agree well with those of Jitthai (2012) who
reported that Urit gave significantly higher Hb values than the automated blood
46
analyser. Generally, the variation of Hb values was slightly higher at low and high
Hb concentrations and least at normal Hb concentration suggesting that majority of
the laboratories may be using calibrators for the normal Hb concentration only while
calibrating the analysers and fail to include calibrators for low and high Hb
concentrations. The results presented in the current study demonstrate that the
problem of inter-laboratory variability even between those laboratories that use the
same brand of analyser is still big suggesting that there are inconsistent calibrations
of instruments at the laboratory level. It is apparent that the variation of Hb
measurements varies across the type of analyser used and the methods.
One possible explanation for deviant results from the different laboratories could
have been due to “matrix effects”. Processed blood haemolysate is prepared in a
matrix that is less complete than patient’s specimens and may have properties that
differ from pure blood. For example, addition of stabilizers or other additives that
are used in the preparation of processed blood samples are known to alter the
properties of the specimen (Rej, Jenny and Bretaudire, 1984., Fasce et al., 1973,
Breutaudire et al., 1981, Ulda, 1984, Fraser & Peake, 1980, Rej & Drake, 1991).
Such anomalies are considered to be “matrix effects” (Clark, Kricka & Whitehead,
1981). Although matrix effects might explain some of the evident method- specific
biases reported in this study, and which is expected to be more or less constant
across all analyzers (Miller & Kaufman, 1993; Power, 1995). The widely variant
results obtained with Sahli, Diaspect, Hemocue, Mission and Urit are less likely to
be caused by matrix effects but may rather reflect serious calibration problems with
these instruments in the field. However, other causes of bias that can result in false
47
test results include instrument calibration errors, reagent lot differences, inaccurate
dilutions, personal performance error and other factors (Klee, 1999; Jenny &
Jackson, 2000; Miller, Levinson & Elin, 2003.
The present study has demonstrated that there is considerable high variation of Hb
values when using manual methods when compared with the automated methods.
All the analysers that produced Hb values that deviated significantly from the
expected mean values employed the manual method of operation. Also as is evident
in this study, the manual methods showed higher CVs than the automated methods
indicating that the automated methods generally have higher precision than the
manual methods. These results agree with those of a previous study conducted by
Fink et al., (1997) where manual methods produced greater CVs than the automated
methods. Also, as reported by Bilto (1999), in his study carried out to assess the
analytical quality of Jordanian haematology laboratories, the cell counter methods
showed better inter-laboratory agreement than the manual methods. Inherent errors
in the manual methods of Hb value estimation on a sample and errors caused by the
observer would be the causes of this great variation seen in the manual methods. Use
of calibration specimens and high quality control Hb solutions; checking and
calibration of equipment are crucial for proper determination of Hb in the blood
(Pupkova et al., 2002). These apparent differences in Hb results obtained by
different laboratories using different analysers/methods (inter-device) and in
different laboratories using the same analysers/methods (intra- device) can be
addressed by harmonization of procedures for Hb measurement in all the
laboratories.
48
The results obtained from the questionnaire indicate that very few (15%)
laboratories participate in External Quality Assessment Schemes/PT programmes
(Table 4-10). A possible explanation to this may be due to lack of funds since EQA
is expensive, lack of awareness by some of the laboratories and lack of commitment
by the laboratory staff and management as well. Regular participation in EQA
assists individual laboratories to continuously monitor their performance and to
compare it with that of other laboratories. Also, it is an effective means for
identifying problems that cause interlaboratory variation of laboratory results and
initiates a process towards solving these problems thus improving the quality of
service at the level of each individual laboratory. Lack of participation in EQA is
one of the factors among others, which cause interlaboratory variation of laboratory
results.
Different laboratories used different reagents and control materials depending on the
method of analysis/analysers used in their laboratory. In order to achieve
standardisation and interlaboratory comparability of Hb results, manufacturers
should consider coming up with a single common type of calibration material that
can be used for all analytical methods and analysers. A total of 55% of laboratories
reported not using quality control/reference materials when performing Hb test. This
could be due to the fact that quality control materials are expensive and are not
readily available to most laboratories. A close observation of the results indicate that
even those laboratories that use quality control materials do not use all the three
levels but only the normal Hb concentration.
49
A good number (31.5%) of the laboratories did not use SOPs when performing Hb
tests, however, most laboratories reported not experiencing any challenges in the
analysis of the samples. These poor quality control practices further increase the
problem of interlaboratory variation of Hb measurements. Recent recommendations
from the European External Quality Assessment–Organizers Working Groups A and
B ( Libeer et al., 1996; Thienpont et al., 1995), the European Union directive for in
vitro diagnostic medical devices (Directive 98/79/EC 1998) the International
Organization for Standardization (ISO 17511:2003, 2007) and the Clinical and
Laboratory Standards Institute (CLSI Document X5-R, 2006) have recognized the
objective of harmonized results and the importance of using commutable reference
materials for method calibration traceability and for proficiency testing (PT)
programs. To ensure standardisation of Hb measurements throughout the world, The
International Council for Standardization in Haematology (ICSH) in conjunction
with Eurotrol, B.V. (Ede, NL; http://www.eurotrol.com) released a new lot of the
haemiglobincyanide (HiCN(Fe), HiCN) standard in 2008 for use in the
standardization and calibration of whole blood Hb measurements on most
haemoglobinometers and automated blood cell counters (Davis & Jungerius, 2010).
However this reference standard is not readily available in the developing countries
laboratories. There was a slight variation in the reference ranges being used in the
laboratories which can have impacts on patient classification.
5.2. Conclusions
Laboratories participating in EQAs of Hb measurements are able to accurately
differentiate low and normal Hb specimens and to a lesser extent, high Hb specimen.
50
Inter-laboratory variability of Hb measurements in an external quality assessment is
quantifiable. Inter-laboratory variation of Hb results for samples with low, normal
and high Hb concentrations was 33.3%, 25.1% and 29.4% respectively.
The results have shown that Celltac, Humalyzer Junior, Medonic, Mindray,
Colourimeter, Hemocontrol and Sysmex produce accurate and reliable Hb results
than Diaspect, Sahli, Hemocue, Urit and Mission.
The manual methods produced higher CVs than the automated methods thus
indicating that manual methods generally have lower precision.
5.3. Recommendations
Clinical laboratories need to regularly participate in continuous EQAS in
order to achieve interlaboratory comparability of Hb results.
There is need for an oversight body and policies to be established by the
government that will oversee the implementation of method comparison
studies in both private and public hospital laboratories in order to ensure
interlaboratory and inter-method comparison of analytical results for Hb.
This study has established the need for validation and standardisation of all
analysers, methods and procedures of Hb measurement being used in the
Kenyan clinical laboratories in order to achieve interlaboratory
comparability.
Laboratories should embrace automation and gradually replace manual
methods with automated methods of Hb measurement which are more
accurate and reliable.
51
REFERENCES
Amukele, T., Michael, K., Hanes, M., Miller, R. & Jackson J. (2012). External quality
assurance performance of clinical research laboratories in Sub-Saharan Africa.
American Journal of Clinical Pathology, 138 (5), 720-723.
Argawal, R. & Heinz, T. (2001). Bedside haemoglobinometry in hemodialysis patients:
Lessons from point of care testing. ASAIO Journal, 47(3), 240-243.
Barbara, H., Anna, P. & Norma, J. (2000). Basic Medical Laboratory Techniques. 4th
Ed. Albany, NY. Delmar Publlishers.
Barduagni, P., Ahmed, A.S., Curtale, F., Raafat, M. & Soliman, L. (2003). Performance
of Sahli and colour scale methods in diagnosing anaemia among school children in
low prevalence areas. Trop Med Int Health, 8 (7), 615 – 618.
Bhaskaram, P., Balakrishna, N., Radhakrishna, K.V. & Krishnaswamy, K. (2003).
Validation of haemoglobin estimation using hemocue. Indian J Pediat, 70, 25-28.
Bilto, Y. Y. (1999). Consensus and accuracy in haematology laboratories of developing
countries: the Jordanian experience. Clin Lab Haematol, 21 (1), 11–15.
Birjees, M.K., Nazish, G.I., Nadira, J., Nadia, N. & Masood, A. (2009). External Quality
Assessment Scheme (IEQAS) in Haematology at National Institute of Health,
Islamabad, Pakistan. J Coll Physicians Surg Pak, 19 (12), 786-790.
Bland, J.M. and Altman, D.G. (1986). Statistical-methods for assessing agreement
between two methods of clinical measurement. The Lancet, 1, 307–310.
Blerk, M.V., Coucke, W., Chatelain, B., Goossens, W., Jochmans, K., Meeus, P… &
Libeer, J-C. (2007). External quality assessment in the measurement of
52
haemoglobin by blood gas analysers in Belgium. Scand J Clin Lab Invest, 67, 735–
740.
Boulton, F., Nightingale, M. & Reynolds, W. (1994). Improved strategy for prospective
blood donors for anemia. Transfusion Medicine, 4, 221-225.
Bretaudire, J-P., Dumont, G., Rej, R., & Bailly, M. (1981). Suitability of control
materials. General principles and methods of investigation. Clin Chem, 27, 798-805.
Burger, S. & Pierre-Louis, J. (2003). A Procedure to Estimate the Accuracy and
Reliability of HemoCue™ Measurementsof Survey Workers. Retrieved from:
[http://www.idpas.org/pdf/2362ProcedureToEstimate.pdf] website [accessed 12th
July 2013].
Carter, J.Y., Lema, O.E., Adhiambo, C.G. & Materu, S.F. (2002). Developing external
quality assessment programmes for primary health: care level in resource
constrained countries. Accredit Qual Assur, 7, 345-350.
Chen, P., Short, T., Leung, D. & Oh, T. (1992). A clinical evaluation of the Hemocue
haemoglobinometer using capillary, venous, and arterial samples. Anaesth Intensive
Care, 20, 497-503.
Clark, P.M., Kricka, L.J. & Whitehead, T.P. (1981). Matrix effects in clinical analysis:
commutability of control materials between Ektachem, Beckman and SMA 12/60
glucose and urea methods. Clin Chim Acta, 113, 293-303.
Clinical and Laboratory Standards Institute. (2006). Metrological traceability and its
implementation: a report, Joint IFCC-CLSI project, CLSI document, EP32-R
53
(Formerly X05-R). CLSI, Wayne, Pennsylvania, USA, 2006. (Sample retrieved
from www.ifcc.org or www.clsi.org)
Clinical Laboratory Improvement Act. (1988). Federal Register, Feb 28, 1992, 57 (40),
7002- 7186.
Cochran, W. G. (1977). Sampling techniques (3rd ed.). New York: John Wiley & Sons.
Croll, M.H., Chesler, R., and Elin, R.J. (1989). Effect of lyophilization on results of five
enzymatic methods for cholesterol. Clin Chem, 35, 1523-1526.
Davis, B.H. & Jungerius, B. (2010). International Council for Standardization in
Haematology technical report 1-2009: new reference material for
haemiglobincyanide for use in standardization of blood haemoglobin measurements.
Int J Lab Hematol, (32), 139–141.
De Louw, A. V., Lasserre, N., Drouhin, F., Thierry, S., Lecuyer, L., Caen, D. &
Tenaillon, A. (2007). Reliability of HemoCue in patients with gastrointestinal
bleeding. Intensive Care Med, 33 (2), 355-358.
Deom, R., El Aouad, C.C., Heuck, S., Kumara, S.M., Lewis, A. & Wardle, J. (1999).
Requirements and Guidance for External quality Assessment Schemes for Health
Laboratories. WHO/DILJLAB / 99.2.
Directive 98/79/EC of the European Parliament and of the Council of 27 October 1998
(1998). In vitro diagnostic medical devices. Off J Eur Communities, L 331. (41), 1–
37.
EAREQAS. ( 2010). Report of the East African Intercountry Meeting on Strengthening
Capacity for Post Market Surveillance of Diagnostics & third regional meeting of
the East African Regional External Quality Assessment Scheme. Nairobi, Kenya.
March 15-17, 2010.
54
Fasce, C.F., Rej, R., Copeland, W.H., & Vanderlinde, R.E. (1973). A discussion of
enzyme reference materials: applications and specifications. Clin Chem, 19, 5-9.
Fierdoz, O. (2012). Essential laboratory knowledge for the clinician. Laboratory testing
forms an integral part of patient management. Contin Med Educ, 30 (7), 244-248.
Fink, N.E., Fernandez Alberti, A. & Mazziota, D. (1997). External assessment of analytic
quality in haematology: a necessity in Latin America. Rev Panam Salud Publica,
2(3); 181-188.
Fraser, C.G. &Peake, M.J. (1980). Problems associated with clinical chemistry quality
control materials. Crit Rev Clin Lab Sci, (12), 59-86.
Frean, J., Perovic, O., Fensham, V., McCarthy, K., Poonsamy, B., Dini, L . . . &
Ndihokubya, J. (2012). External quality assurance of national public health
laboratories in Africa, 2002-2009 [abstract]. Bull World Health Organ, 90, 191-199.
Gehring, H., Duembgen, L., Peterlein, M., Hagelberg, S. & Dibbelt, L. (2007).
Hemoximetry as the “gold standard”? Error assessment based on differences among
identical blood gas analyzer devices of five manufacturers. Anesth Analg, 105, 24–
30.
Gomez, S.A., Navarro, N.L., Pérez, C.E., Lozano, M.L., Candela, J.M., Cascales, A . . .
& Rivera, J. (2007). Evaluation of four rapid methods for hemoglobin screening of
whole blood donors in mobile collection settings. Transfus Apher Sci, 36 (3), 235-
242.
International Committee for Standardization in Haematology—Expert Panel on
Haemoglobinometry. (1987). Recommendations for reference method for
haemoglobinometry in human blood (ICSH standard 1986) and specifications for
55
international haemoglobincyanide reference preparation (3rd Ed.). Clin Lab
Haematol, 9, 73–79.
ISO 17511:2003. In Vitro Diagnostic Medical Devices—Measurement of Quantities in
Biological Samples—Metrological Traceability of Values Assigned to Calibrators
and Control Materials. Geneva, Switzerland: International Organization for
Standardization. Available at: http://www.iso.org/. [accessed October 28, 2007].
Jenny, R.W., & Jackson-Tarentino, K.Y. (2000). Causes of unsatisfactory performance in
proficiency testing. Clin Chem, 46 (1), 89-99.
Jitthai, S. (2012). A comparison of haemoglobin measurement between
haemoglobinometer (POCT) and automated blood cell analyzer in hemodialysis
patient. J Med Technol Assoc of Thai, 40 (3), 4332 - 4337.
Kapil, U., Tandon, M. Pathak, P. & Dwivedi, S.N. (2002). Comparison of haemoglobin
values obtained by Haemocue and Sahli’s methods. Ind J Public Health, 46 (1), 28 -
30.
Klee, G.G. (1995). Analytic performance goals based on direct effect of analytic bias on
medical classification decisions. CDC, 1995 Institute: Frontiers in Laboratory
Practice Research, 219–226.
Klee, G.G. (1999). Maximizing efficacy of endocrine tests: importance of decision-focused
testing strategies and appropriate patient preparation. Clin Chem, 45 (8(B)), 1323-
1330.
Kaplan, L.A., Pesce, A.J. & Kazmierczak, S.C. (ed.) (2003). Clinical Chemistry: Theory,
Analysis, Correlation. 4th Ed. St Louis: CV Mosby.
Kisabei, P. (2013). KMLTTB registered laboratories. [E-mail]. Message to: Mandania, E.
13th June, 2013.
56
Lauren, B. (2013). Factors affecting haemoglobin measurement. [Online] DOI
10.1007/s10877-013-9456-3.Retrieved from:
http://link.springer.com/article/10.1007%2Fs10877-013-9456-3. [Accessed 12th
may 2013].
Lauren, B., Stephanie, R. & Erin, M. (2011). Continuous noninvasive haemoglobin
monitoring during complex spine surgery: IARS, 113 (6), 1396-1402.
Lewis, S.M., (1998). Quality assurance in haematology. World Health Organization.
(WHO/lab/98.4), 1998: p.104 published online at:
http://whqlibdoc.who.int/hq/1998/WHO_LAB_98.4.pdf. [accessed 12th July 2013].
Libeer, J.C., Baadenhuijsen, H., Fraser, C.G., Petersen, P.H., Ricoss, C., Stock, D. &
Thienpont, L. (1996). Characterization and classification of external quality
assessment schemes (EQA) according to objectives such as evaluation of method
and participant bias and standard deviation. Eur J Clin Chem Clin Biochem, 34,
665–678.
Lippi, G. & Guidi, G.C. (2007). Risk management in the preanalytical phase of laboratory
testing. Clin Chem Lab Med, 45, 720 – 727.
Lippi, G., Banfi, G., Buttarello, M., Ceriotti, F., Daves, M. & Dolci, A. (2007).
Recommendations for detection and management of unsuitable samples in clinical
laboratories. Clin Chem Lab Med, 45, 728 – 736.
Looker, A. C., Sempos, CT., Liu, K. A., Johnson, C. L. & Gunter, E. (1990). Within
person variance in biochemical indicators of iron status: Effects on prevalence
indicators. Amer J Clin Nutr, 52, 541-547.
Martin, D. T., Ashenden, M., Parisotto, R., Pyne, D. & Hahn, A. G. (1997). Blood
testing for professional cyclists: what's a fair hematocrit limit? Sportscience News.
57
[Retrieved from] http://www.sportsci.org/news/news9703/AISblood.html. [accessed
27th March 2013].
Masimo Cooperation: Technical Bulletin. (2009). Total Haemoglobin Measurements:
Accuracy of Laboratory Devices and Impact of Physiologic Variation. Supplement
S26-030F 7799-5447A-0109. Retrieved from:
[http://www.masimo.com/pdf/SpHb/LAB5447A.pdf] [accessed 12th April 2013].
Miller, W.G. & Kaufman, H.W. (1993). College of American Pathologists Conference
XXIII on matrix effects and accuracy assessment in clinical chemistry: introduction.
Arch Pathol Lab Med, 17, 343-344.
Miller, J.J., Levinson, S.S., Elin, R.J. (2003). Interferences in laboratory tests. In: Clinical
diagnostic technology: the total testing process. Volume 2: the analytical phase.
Washington, DC: American Association of Clinical Chemistry, 2003.
Mokken, F. C., Van der, F. J., Henny, C. P., Goe dhart, P. T. & Gelb, A. W. (1996).
Differences in peripheral arterial and venous hemorheologic parameters. Annals of
Hematology, 73 (3), 135-137.
Mohanram, M., Ramana-Rao G.V. & Sastry J.G. (2002). A comparative study on
prevalence of anaemia in women by cyanmethemoglobin and hemocue methods.
Indian J Community Med, 27 (2), 58-61.
Morris, S., Ruel, M., Cohen, R., Dewey, K., Brière, B. D. L. & Hassan, M. (1999).
Precision, accuracy, and reliability of haemoglobin assessment with use of capillary
blood. Am J Clin Nutr, 69 (6), 1243-1248.
Munich, D.S. (1982). A New Concept for Quality Control of Clinical Laboratory
Investigations in the Light of Clinical Requirements and Based on Reference
Method Values. J. Clin. Chem. Clin. Biochem, 20, 817-824.
58
Neufeld, L., Garcia-Guerra, A., Sanchez-Francia, D., Newton-Sanchez, O., Ramirez-
Villalobos, M.D. & Rivera-Dommarco, J. (2002). Haemoglobin measured by
hemocue and a reference method in venous and capillary blood: a validation study.
Salud Publica Mex, 44, 219 – 227.
Olafsdottir, E., Hellsing, K., Steensland, H., Tenhunen, R. & Uldall, A. (1994). Adding
new scopes of traditional EQA schemes emphasizing quality improvement. Upsala J
Med Sci, 99; 377-384.
Patel, K.P., Hay, G.W., Cheteri, M.K. & Holt, D.W. (2007). Haemoglobin test results
variability and cost analysis of eight different analyzers during open heart surgery. J
Extra Corpor Technol, 39, 10-17.
Patil, P., Thakare, G. & Patil, S. (2013). Variability and Accuracy of Sahli’s Method in
Estimation of Haemoglobin Concentration. Natl J Integr Res Med, 4 (1), 38 – 44.
Plebani, M. (2009). Exploring the iceberg of errors in laboratory medicine. Clin Chim Acta,
404; 16 – 23.
Plebani, M. (2010). The detection and prevention of errors in laboratory medicine. Ann
Clin Biochem, 47, 101 – 110.
Plebani, M. & Lippi, G. (2010). To err is human. To misdiagnose might be deadly. Clin
Biochem, 43, 1 – 3.
Power, D.M. (1995). Overcoming the limitations of artificial clinical specimens: linkage to
performance of patients’’ specimens [editorial]. Clin Chem, 41, 1689-1692.
Prakash, S., Kapil, U., Singh, G., Dwivedi, S.N & Tandon, M. (1999). Utility of
hemocue in estimation of haemoglobin against standard Blood Cell Counter method.
J Assoc Physicians India, 47 (10), 995-997.
59
Pupkova, V.I., Zhdanova, V.V., Stepanova, E.G., Leonova, T.I. & Palamarchuk,
M.B. (2002). Quality problems in quantitative determination of Hb in blood. Klin.
Lab. Diagn, 12, 36-39.
Rechner, I.J., Twigg, A., Davies, A.F. & Imong, S. (2002). Evaluation of the hemocue
compared with the coulter STKS for measurement of neonatal haemoglobin. Arch
Dis Child Fetal Neonatal Ed, 86; F188 – 189.
Rej, R. & Drake, P. (1991). The nature of calibrators in immunoassays: are they
commutable with test samples? Must they be? Scand J Clin Lab Invest, 51 (205),
47-54.
Rej, R. & Jenny, R.W. (1992). How good are clinical laboratories? An assessment of
current performance. Clin Chem, 38(7), 1210-1217.
Rej, R., Jenny, R.W & Bretaudire, J-P. (1984). Quality control in clinical chemistry:
characterization of reference materials. Talanta, 31, 851-862.
Renu, S., Katoch, S. C., Upendra, S., Seema, R. & Hema, A. (2007). Impact of external
haematology proficiency testing programme on quality of laboratories. Indian J Med
Res, 126, 428-432.
Ricos, C., Alvarez, V., Cava, F., Garcia-Lario, J. V., Hernandez, A., Jimenez, C. V . . .
& Simon, M. (1999). Desirable specifications for imprecision, inaccuracy, and total
allowable error, calculated from data on within-subject and between-subject
biologic variation. Available from: [http://www.westgard.com/biodatabase1.htm]
[accessed 19th July 2013].
Ricos, C., Baadenhuijsen, H., Libeer, C.J., Petersen, P.H., Stockl, D., Thienpont, L. &
Fraser, C.C. (1996). External quality assessment: currently used criteria for
60
evaluating performance in European countries, and criteria for future harmonization.
Eur J Clin Chem Clin Biochem, 34, 159-165.
Rippey, J.H. & Williamson, W.E. (1988). The overall role of a proficiency testing
programme. Arch Pathol Lab Med, 112, 340–342.
RNA Medical. (2011). CVC 223 CO-Oximeter Calibration Verification Controls. RNA
Medical, Division of Bionostics. Devens, MA 01434. [Available from]:
http://www.rnamedical.com/pdf/datasheets/cvc223_rev_c.pdf. [accessed 10th July
2013].
Robertson, L., Lewis, D. & Osei-Bimpong, A. (2011). Utility of the Diaspect
Haemoglobinometry System as Point- Of- Care Analyzer. J Near Patient Test
Technol, 10 (1), 17 – 21.
Sah, S.P., Raj, G.A. & Prakash, M.B. (1999). Quality assurance programme in
haematology at a teaching hospital in the eastern region of Nepal. Indian J Pathol
Microbiol, 42 (2), 145-9.
Sari, M., De Pee, S., Martini, E., Herman, S., Sugiatmi. & Bloem M.W, (2001).
Estimating the prevalence of anaemia: a comparison of three methods. Bull World
Health Organ, 79, 506 – 511.
Savage, R. A. (1989). Proficiency testing and laboratory quality. Arch Pathol Lab Med,
113, 983–984.
The Royal College of Pathologists of Australasia. (2010). Chain of Information Custody
for the Pathology Request-Test- Report Cycle in Australia (Guidelines for Pathology
Requesters and Pathology Providers). K:\2002 onward\Operations\Policy Manual
\Guidelines\Chain of Information Custody.DOC. Retieved from:
61
http://www.health.gov.au/internet/main/publishing.nsf/Content/FDD654D0C46DC3
74CA2573A0000A9C1A/$File/ChainInfoCustody.pdf [accessed 17th July 2013].
Thienpont, L., Franzini, C., Kratochvila, J., Middle, J., Ricos, C., Siehnann, L. &
Stockt, D. (1995). Analytical quality specifications for reference methods and
operating specifications for networks of reference laboratories. Eur J Clin Chem
Clin Biochem, (33), 949–957.
Uldal, A. (1984). Apparent interlaboratory variation using different types of reference sera.
Scand J Clin Lab Invest, 44(172), 163-171.
Von Schenck, H., Falkensson, M. & Lundergerg, B. (1986). Evaluation of “hemocue,” a
new device for determining haemoglobin. Clin Chem, 32, 526 – 529.
Westgard J.O. (2008). Use and interpretation of common statistical tests in method
comparison studies. Clin Chem, 54, 612.
World Health Organization. (2008). Worldwide prevalence of anaemia 1993-2005: WHO
global database on anaemia. [Online]. Retrieved from:
[http://whqlibdoc.who.int/publications/2008/9789241596657_eng.pdf] website
[accessed 12th July 2013].
WHO Regional Office for Africa. (2008). Strengthening public health laboratories in the
WHO African region: a critical need for disease control. Retrieved from:
http://www.afro.who.int/en/fifty-eight-session.html. Accessed september 6, 2012.
World Health Organization. (2007). External quality assessment scheme for haematology.
[Internet]. [Cited 2007]. Retrived from:
http://www.who.int/diagnostics_laboratory/quality/haematology/en/
62
Zigilstra, W.G. (1997). Standardization of haemoglobinometry: History and new
challenge. Comp Haematol Int, 1, 125-132.
63
APPENDICES
Appendix I: Procedure for Haemolysate Preparation
STANDARD OPERATING PROCEDURE FOR PREPARATION OF
HAEMOLYSATE SAMPLE
1. Centrifuge anticoagulated blood from a blood bank, with negative serology
for the human immunodeficiency (HIV) hepatitis B (HBsAg), hepatitis C
(HCV) virus in screw capped bottles of appropriate size. Aseptically separate
the plasma from the white layer (layer of leukocytes or buffy coat).
2. Add to each red cell deposit an excess of physiological saline 9g/L (NaCl),
mix well and centrifuge. Discard the supernatant and any remaining buffy
coat.
3. Repeat saline wash two times with sterile physiological saline solution to the
complete elimination of the plasma, leukocytes, and platelets, each time
removing the top layer of the package of erythrocytes in each wash.
4. Add to the washed cells an equal volume of water and 0.5 volumes of
carbon tetrachloride (CCl3). Cap the containers and shake them vigorously in
a mechanical Shaker for 1 hour. Refrigerate overnight to allow the lipids/cell
debris to form a semi-solid interface between carbon tetrachloride and lysate.
5. On the following day, centrifuge at 2 500 g for 20 minutes; transfer the
upper layers into universal containers and centrifuge at about 3000g for one
hour. Collect the upper 95% and pool into a clean tube.
6. To each 70 mL of lysate, add 30 mL of glycerol. Then add antibiotics (25-
50 mg of penicillin) and 25-50 mg of gentamicin per 500 mL of material.
64
This material is stored at 4 ° c until use for short periods or frozen at -20 for
longer period until required for dispensing.
7. If a lower concentration is required, add an appropriate volume of 30% (v/v)
glycerol in a solution of 9 g/l of sodium chloride (NaCl) to the stock. Mix
well.
8. With continuous stirring distribute the mixture aseptically into sterile
containers. Cover and seal.
9. Assign the haemoglobin value. Preserved at 4 ° C, the product should keep
the value that has been assigned for several months.
65
Appendix II: Standard Operating Procedure for Triple Packaging System
STANDARD OPERATING PROCEDURE ON TRIPLE PACKAGING
SYSTEM
The triple packaging system consists of three layers: the primary receptacle, the
secondary packaging and the outer packaging. This packaging system is meant to
withstand leakage of contents, shocks, temperature, pressure changes and other
conditions that can occur during ordinary handling in transportation.
1) Put the sample in a primary receptacle (tubes/vials) and ensure that all the
tubes/vials that are being sent are tightly closed to avoid any leakage of the
sample. The primary receptacle containing the specimen must be watertight,
leak proof and appropriately labelled as to content.
2) Wrap every tube/vial in enough absorbent material like paper towels or
cotton wool to absorb all fluid in case of breakage or leakage.
3) Place the samples in a second watertight, leak proof packaging such as a zip
lock bag so as to enclose and protect the primary receptacle(s). Several
wrapped primary receptacles may be placed in a single secondary packaging.
4) Get a plastic container and place the samples in the zip lock bag into it and
close it.
5) Place four frozen ice packs from the -20 degrees Celsius compartment of the
refrigerator, on the sides of a small cool box ensuring that the secondary
container is in the middle of the ice packs.
6) Place absorbent material on top and if necessary between the ice packs.
66
7) Place the case investigation forms or appropriate laboratory request forms in
a plastic bag to keep from becoming contaminated or destroyed by the wet
ice packs.
Note: An itemized list of package contents must be included between the
outer and secondary container.
8) The cool box is then sealed and addressed to the receiver.
9) Label the box with biohazard symbol and with upward arrows
Note: The outer package must be of adequate strength for its capacity, mass, and
intended use. It must be must be rigid and have one side that is at least 100 mm
X 100 mm, in order for required markings and labels to fit.
67
Appendix III: Instructions for the Participating Laboratories
INSTRUCTIONS FOR HAEMOGLOBIN TEST SAMPLE HANDLING FOR
EQAS IN HB MEASUREMENT
Inter-laboratory variability of haemoglobin measurements obtained from
selected clinical laboratories in Kenya
Introduction
This is a set of instructions for the above study that should be followed by every
participating laboratory so as to ensure proper sample analysis is done.
You are kindly asked to read keenly the instructions below before performing any
analysis of the samples. Feel free to seek any clarification from the Principle
Investigator.
Instructions to be followed before and during sample analysis:
1) Always keep the samples refrigerated at 2 – 6°c when not in use. Do not
freeze the samples.
2) Perform sample analysis within two days after the delivery of the samples to
your laboratory.
3) Verify that the samples are in good condition and well labeled before the
analysis. There are three samples labeled sample A, Sample B & Sample C.
Each vial contains exactly 1ml of sample.
4) Let the sample attain room temperature before carrying out the test.
5) Mix the samples well before the test by gently inverting the vial about 3 – 4
times.
68
6) Make sure that your machine/ equipment is in good working condition
before the analysis.
7) Perform a single haemoglobin test of each of the samples in the morning
hours using the analytical method that is currently being used in your
laboratory. Record the results in the worksheet provided.
8) Perform the second haemoglobin test of each of the samples in the afternoon
hours. Record the results in the worksheet provided.
NOTE: The time difference between the morning and afternoon readings
should not exceed 6 hours.
9) Lastly fill the questionnaire provided.
10) The results and the questionnaire will be collected from the laboratory within
one week of sample delivery.
69
Appendix IV: Questionnaire
QUESTIONNAIRE FOR INTER-LABORATORY VARIABILITY OF
HAEMOGLOBIN MEASUREMENTS STUDY
I am Esther Wangui Mandania, a master’s student at Jomo Kenyatta University of
Agriculture and Technology, faculty of COHES, Department of Medical Laboratory
Sciences. I am conducting research on interlaboratory variability of haemoglobin
measurements obtained from selected clinical laboratories in Kenya.
This is a qualitative questionnaire that is intended to provide the researcher with
important information regarding haemoglobin measurement in the clinical
laboratories of Kenya in order to assess the inter-laboratory variability of
haemoglobin measurements.
Kindly provide the researcher with the following information;
1. Details for the laboratory:
Code assigned to the laboratory (by PI) ------------------------------------------
Mobile Phone No: --------------------------------------------------------------------
Physical Address: P.O BOX--------------------------------------------------------
Email Address------------------------------------------------------------------------
2. Which date and time were the samples received in your laboratory?
Date--------------------------------------Time-------------------------------------
3. In which condition were the samples received by your laboratory?
□ Intact (No damage/breakage)
□ Damaged and leaking
□ Labelled
□ Unlabeled
70
4. What volume of each EQA sample did you receive?
□Less than 0.5mL
□ 0.5mL
□ More than 0.5mL
5. Which equipment and method of haemoglobin estimation was used to
measure haemoglobin of the EQA samples in your laboratory?
Method?--------------------------------------------------------------------------------
Equipment? --------------------------------------------------------------------- ---
6. Please specify the reference levels for haemoglobin used in your laboratory
for the following;
Adult women---------------------------------------------
Adult men-------------------------------------------------
Children----------------------------------------------------
7. Which reagents did your laboratory use for haemoglobin estimation for the
samples? ---------------------------------------------------------------------------------
---------------------------------------------------------------------------------------------
What challenges and difficulties did your laboratory experience in the
analysis of the samples?---------------------------------------------------------------
---------------------------------------------------------------------------------------------
---------------------------------------------------------------------------------------------
State the SOP that your laboratory used during samples analysis? -------------
---------------------------------------------------------------------------------------------
71
8. Please state the control materials that were used during samples analysis -----
---------------------------------------------------------------------------------------------
------------------------------------------------------------- ---------------------------
9. Which EQA scheme for haemoglobin has your laboratory participated in
before? -----------------------------------------------------------------------------------
---------------------------------------------------------------------------------------------
72
Appendix V: Worksheet
Worksheet for inter-laboratory variability of Hb measurements study
Code of laboratory:
Date & time of analysis:
Results (g/dL)
Type of EQA Sample Reading 1 Reading 2
EQA Sample A
EQA Sample B
EQA Sample C
73
Appendix VI: Ethical Clearance
74
75
Appendix VII: Informed Consent Form for Laboratory Management
INFORMED CONSENT FORM
Inter-laboratory variability of haemoglobin measurements obtained from
selected clinical laboratories in Kenya
Introduction
This informed consent form is for the laboratory in-charges, supervisors or
managers of clinical laboratories that i am inviting to participate in research on
haemoglobin measurement. The title of the research is “Inter-laboratory
variability of haemoglobin measurements obtained from selected clinical
laboratories in Kenya”
Principal investigator:
Name: Esther Wangui Mandania
Study Authority: Jomo Kenyatta University of Agriculture and Technology
Faculty: College of Health Sciences
Department: Medical Laboratory Sciences
Address: P.O Box 2028, Thika
Phone No: 0727906351
E-mail: [email protected]
Authorizing body: KNH/UON-ERC
This Informed Consent Form has two parts:
Information sheet (to share information about research with you)
Certificate of consent (for signatures if you agree to take part)
76
You will be given a copy of the full informed consent form
PART I: Information Sheet
I am Esther Wangui Mandania, a master’s student at Jomo Kenyatta University of
Agriculture and Technology, faculty of COHES, Department of Medical Laboratory
Sciences.
I am doing research on variability of haemoglobin measurements between clinical
laboratories in Kenya. I am going to give you information and invite you to
participate in this research. Before you decide to participate in this research, you can
talk to anyone you feel comfortable with about the research. Also ask any questions
that you may have about the research. You do not have to decide today whether or
not you will participate in the research, sufficient time will be given to you to make
this decision.
If there is anything or any word that you do not understand, please ask me to stop as
we go through the information and i will take time to explain.
Purpose of the research
Comparability of haemoglobin test results is difficult due to variability of
haemoglobin measurements between laboratories. Laboratories should strive to
produce reliable results that are consistent between laboratories so that comparable
results for the same test are obtained in all the laboratories. The reason i am doing
this research is to find out if laboratories can give the same results for the same test
and to determine the variability of haemoglobin measurements between clinical
laboratories in Kenya. The information obtained from the research will help to
identify areas of improvement in haemoglobin measurements in the laboratories.
77
Procedures
This research will involve analysis of three EQA samples only (A, B and C)
which will be analysed the same day and filling the results of analysis in a
worksheet.
If your laboratory agrees to participate in this research you will receive three
EQA samples labelled A, B and C accompanied by a questionnaire and a
worksheet.
I will request you to analyse the samples within two days in the routine way
as patient samples are analysed, record the results in the worksheet provided
and fill the questionnaire.
You will then send the results and the questionnaire to the principal
investigator or they will be picked from the laboratory whichever you prefer.
Duration
Your expected time commitment for this study will not be more than four days in
total. However the investigator will take about 6 months to complete the whole
research study.
Risks
There are no any risks for your participation in this study.
Benefits
If you (your Laboratory) participates in this research, you will receive a summary of
your(laboratory’s) performance which will include the results from the laboratory,
the expected values (targets) and the potential causes of errors and remedial actions(
for laboratories that will not be in a position to give the target values). This will
provide an opportunity to the laboratory to identify the areas of improvement in the
78
measurement of haemoglobin as well as in general areas of your laboratory, which
the laboratory can work on to improve the quality of the results.
It is hoped that your participation will help the researcher determine and learn more
about variability of haemoglobin measurements between clinical laboratories in
Kenya, how they impact on patient management and how inter-laboratory
comparability of results can be achieved.
Confidentiality
The information that will be collected from this research project will be kept
confidential. The results from your laboratory and all information about the
laboratory will not be disclosed to any other laboratory or party. The results from
your laboratory and any information about your laboratory will have a number on it
instead of the name of the laboratory. Only the researcher will know what your
laboratory number is. During dissemination of the research findings, this
information will only be reported as group data with no identifying information. All
data, including questionnaires and worksheets for the study will be kept in a secure
location under lock and key and no one will have access to them except the
researcher. After the research is complete, all the raw data that is, the questionnaires
and worksheets will be destroyed by shredding followed by burning.
Compensation
You will not be given any compensation to take part in this research.
Voluntary Participation
Your Participation in this research study is entirely voluntary. It is your choice to
decide whether to participate or not. Whether you choose to participate or not, this
will not affect the relationship you have with the researcher. You may change your
79
mind later and stop participating even if you agreed earlier. Also you have the right
to withdraw at anytime or refuse to participate entirely in the research study without
any consequences. If you do decide to take part in this study, you will be asked to
sign the consent form.
Dissemination of findings
Research findings will be disseminated through publications, conferences and
through other scientific reports however, no identifying information will be used so
as to maintain confidentiality and privacy. Also you will receive a summarised
report of your laboratory’s performance.
Who to contact
If you have any questions regarding this study, you may ask them now or later, even
after the study has started. If you wish to ask questions later, you may contact any of
the following:
Principal Investigator:
Name: Esther Wangui Mandania
P.O BOX 2028, Thika
Mobile telephone: 0727-906351 or 0789-641355
E- Mail address: [email protected].
Supervisor:
Dr. Juliette R. Ongus
Jomo Kenyattta University of Agriculture and Technology
Department of Medical Laboratory Sciences
Mobile Telephone: +254 722339682
80
E- Mail address: [email protected]
KNH/UON-ERC Secretariat:
PROF. M L. CHINDIA
SECRETARY, KNH/UON-ERC
Tel: 726300-9 Ext. 44102 OR +2542726300-19 Ext.44102
Post address: P O BOX 20723-00202, Nairobi, Kenya
Physical Address: School Of Pharmacy, UON behind KNH Dental clinic
E-mail: [email protected]
Website: www.ounbi.ac.ke
Link: www.uonbi.ac.ke/activities/KNHUoN
This proposal has been reviewed and approved by the KNH/UON-ERC. If you wish
to find out more, contact the KNH/UON-ERC secretariat using the contact provided
above.
PART II: Certificate of Consent
I have read the foregoing information and understood, i have had the opportunity to
ask questions about it and any questions that i have asked have been answered to my
satisfaction. I understand that my participation is voluntary and that i am free to
withdraw at any time without giving any reason and without cost. I have received a
copy of this consent form. I voluntarily agree to participate in this study.
Laboratory Code: --------------------------------------------------------------------------------
81
Name of laboratory in-charge/supervisor/manager: ----------------------------------------
Signature: -------------------------------------------------------------Date: --------------------
Name of PI: ----------------------------------------------------------- ID No: -----------------
Signature: ------------------------------------------------------------- Date: -------------------
